06 Fakultät Luft- und Raumfahrttechnik und Geodäsie
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/7
Browse
Search Results
Item Open Access Nonlinear model predictive control of floating wind turbines(2013) Schlipf, David; Sandner, Frank; Raach, Steffen; Matha, Denis; Cheng, Po WenIn this work a nonlinear model predictive control method for a floating wind turbine is presented. A reduced nonlinear model including disturbance preview of wind and waves is derived and implemented to compute optimal input trajectories for collective pitch and the generator torque. A cost functional is introduced which fulfills all desired constraints and controller goals for above rated wind conditions. The controller is tested for extreme and fatigue load cases and a significant reduction of the power and rotor speed deviations is obtained. Furthermore, ultimate tower loads and damage equivalent loads on shaft and blades are decreased. Although more detailed testing is necessary, this preliminary results show the advantages of nonlinear model predictive control for floating wind turbines.Item Open Access Direct speed control using LIDAR and turbine data(2013) Schlipf, David; Fleming, Paul; Kapp, Stefan; Scholbrock, Andrew; Haizmann, Florian; Belen, Fred; Wright, Alan; Cheng, Po WenLIDAR systems are able to provide preview information of the wind speed in front of wind turbines. One proposed use of this information is to increase the energy capture of the turbine by adjusting the rotor speed directly to maintain operation at the optimal tip-speed ratio, a technique referred to as Direct Speed Control (DSC). Previous work has indicated that for large turbines the marginal benefit of the direct speed controller in terms of increased power does not compensate for the increase of the shaft loads. However, the technique has not yet been adequately tested to make this determination conclusively. Further, it is possible that applying DSC to smaller turbines could be worthwhile because of the higher rotor speed fluctuations and the small rotor inertia. This paper extends the previous work on direct speed controllers. A DSC is developed for a 600 kW experimental turbine and is evaluated theoretically and in simulation. Because the actual turbine has a mounted LIDAR, data collected from the turbine and LIDAR during operation are used to perform a hybrid simulation. This technique allows a realistic simulation to be performed, which provides good agreement with theoretical predictions.Item Open Access Application of spaceborne geodetic sensors for hydrology(2013) Tourian, Mohammad J.; Sneeuw, Nico (Prof. Dr.-Ing.)How much freshwater do we have on land? How is the freshwater cycle changing with time? Actually, we can not properly answer these questions as our knowledge of the spatial and temporal dynamics of the hydrological cycle is limited. The lack of knowledge is mainly induced by shortage of observational evidence, which motivates the objective of this study: the monitoring of the hydrological cycle using spaceborne geodetic sensors. Among the current space geodetic sensors, GRACE and satellite altimetry are the two active mission concepts, that can capture part of the hydrological cycle. However, monitoring the hydrological cycle using these two sensors is challenging. Satellite altimetry is investigated as an independent spaceborne sensor that provides the water level and discharge time series. An algorithm is developed to improve the quality of water level time series over inland water surfaces. This algorithm particularly deals with the challenges of resolution and uncertainty of altimetry. The obtained altimetric water level time series is validated against in situ measurements showing about 10% improvement in accuracy of the time series. Moreover, this study proposes an algorithm to reduce the random noise from pre-retracked data. The algorithm combines the results of different retrackers and provides water level time series with reduced noise level. The validation shows a significant reduction of noise level and a clear improvement in correlation with in situ measurements. Moreover, this study proposes a statistical approach based on quantile functions to infer a functional relationship between altimetric water level and in situ river discharge without the need for synchronous data sets. This method is based on a scatter diagram of quantile functions, in which the probability-coordinate is eliminated. In contrast, the conventional methods for simultaneous measurements operate directly on time series and eliminate the time-coordinate. The results show that the proposed methodology provides the same range of error as the common conventional empirical method. The good performance of the statistical approach supports the usage of altimetry to salvage pre-satellite altimetry discharge data and turn them into active use for the satellite altimetry time frame. In addition, a stochastic process model is implemented to (i) deal with the data outages in altimetric discharge, (ii) provide a scheme for data assimilation and (iii) smooth the discharge estimation. The model benefits from the cyclostationary behaviour of the discharge and is combined with the estimated discharge from altimetry and available in situ measurements to form a linear dynamic system. The dynamic system is solved using the Kalman filter, that provides an unbiased discharge with minimum variance. The error level of the results is comparable to the empirical approach. In this study, the utility of GRACE data as sensor of hydrological water storage changes is shown to be limited by the following challenges: consistency, resolution, separability and uncertainty. The challenge of inconsistency is addressed by developing two filters for hydrological and hydro-meteorological water storage changes, which lead to a better correlation with GRACE mass storage changes. The challenges of separability and resolution are not specifically investigated in this study, yet their consequences, which appear in different forms of uncertainties is investigated. To deal with the GRACE uncertainties, an algorithm is developed to detect outliers in monthly solutions. The outliers have been corrected by replacing them by an inter-annual monthly mean of the respective month. The results conclude that outlier identification and correction must be performed before further assimilation of GRACE products into hydrological or hydro-meteorological analysis. Further, a longrange correlation has been identified as another source of uncertainty in GRACE monthly solutions. EOF analysis is employed to identify the zonal behaviour of the GRACE C20 errors as the responsible source for the long-range correlation. It is considered as an error source because its residual contains tidal aliasing errors instead of white noise. Therefore, to reduce the uncertainties in GRACE monthly solutions, tidal aliasing errors are also investigated. Primary and secondary tidal aliasing errors of main tidal constituents, S1, S2, P1, K1, K2, M2, O2, O1 and Q1 are identified in GRACE monthly solutions. The effect of tidal aliasing error is estimated using a least squares Fourier analysis indicating errors up to 22mm over the globe. In general, after dealing with GRACE’s challenges and achieving a data set without outliers, long-range correlation and tidal aliasing errors, the noise level of GRACE is quantified. The quantification shows a variation between 2–20mm/month over different parts of the globe, with higher values over tropical and boreal regions. The results specifically confirm that small catchments in the tropics contain more noise contamination. It is also shown that a lower noise level of a catchment does not necessarily lead to a better correlation of GRACE with hydro-meteorological signal. Finally, the joint performance of spaceborne geodetic sensors for estimating the actual evapotranspiration ETa is assessed. There, two approaches are introduced to estimate ETa using the results of GRACE and satellite altimetry. The results of both approaches are compared with different models and their ensemble mean. All in all, given the obtained relative discrepancy, the methods seem to be a viable way for determining ETa for most non-desert catchments containing hot and warm summers.Item Open Access Adaptive Vorsteuerung für Windenergieanlagen(2013) Schlipf, David; Cheng, Po WenDer Beitrag beschreibt, wie Windmessungen mit LIDAR in einer Vorsteuerung verwendet werden können, um die Drehzahlschwankungen und damit die Belastungen von Windenergieanlagen zu reduzieren. Kernstück dieser Vorsteuerung ist ein Filter, der adaptiv auf die aktuellen Messungen eingestellt werden muss, da sich die Prädiktionszeit und die Korrelation zwischen Vorhersage und Anlagenverhalten kontinuierlich ändern. Die Ergebnisse werden mit Messdaten einer 5MW Anlage validiert.Item Open Access Lidars and wind turbine control. Pt. 1(2013) Schlipf, DavidIn recent years lidar technology found its way into wind energy. The main application is still the site assessment, but the possibility to optimize the energy production and reduce the loads by nacelle or spinner based lidar systems is becoming an important issue. In terms of control the inflowing wind field is the main disturbance to the wind turbine and most of the wind turbine control is designed to deal with variations in this disturbance. From control theory, the control performance can be improved with the knowledge of the disturbance. Due to the measurement principle and the complexity of the wind lidar assisted control is a wide field of research. The main idea is to divide the problem in a measurement and a control problem. The presented work describes first how wind characteristics, such as wind speed, direction and shears, can be reconstructed from the limited provided information (see Section 9.2). Based on the models of the wind turbines (see Section 9.3) it is investigated in Section 9.4, how well the lidar information can be correlated to the turbines reaction. In the next sections, several controllers are presented, see Table 15. All controllers are designed first for the case of perfect measurement and then adjusted for realistic measurements. The most promising approach is the collective pitch feedforward controller using the knowledge of the incoming wind speed providing an additional control update to assist common collective pitch control. Additional load reduction compared to the sophisticated feedback controllers could be archived (Schlipf et al., 2010a). The concept has been successfully tested on two research wind turbines (Schlipf et al., 2012a; Scholbrock et al., 2013). Then a feedforward control strategy to increase the energy production by tracking optimal inflow conditions is presented. The comparison to existing indirect speed control strategies shows a marginal increase in energy output at the expense of raised fluctuations of the generator torque (Schlipf et al., 2011). A Nonlinear Model Predictive Control (NMPC) is also presented, which predicts and optimizes the future behavior of a wind turbine using the wind speed preview adjusting simultaneously the pitch angle and the generator torque. The NMPC achieves further load reductions especially for wind conditions near rated wind speed (Schlipf et al., 2012b). Furthermore, a cyclic pitch feedforward controller using the measured horizontal and vertical shear is introduced to assist common cyclic pitch control for further reduction of blade loads. Simulations results from Dunne et al. (2012) are promising, but they have to be further investigated under more realistic conditions. Finally, the benefit of lidar assisted yaw control is explored. A promising way to obtain a accurate measurement of the wind direction is to measure it over the full rotor plane ahead of the turbine by lidar. The expected increase of the energy output is about one percent of the annual energy production, when using the wind direction signal from the lidar system instead of the sonic anemometer (Schlipf et al., 2011).