Improving Site-Dependent Wind Turbine Performance Prediction Accuracy Using Machine Learning

Data-driven wind turbine performance predictions, such as power and loads, are important for planning and operation. Current methods do not take site-specific conditions such as turbulence intensity and shear into account, which could result in errors of up to 10%. In this work, four different machine learning models (k-nearest neighbors regression, random forest regression, extreme gradient boosting regression and artificial neural networks (ANN) are trained and tested, firstly on a simulation dataset and then on a real dataset. It is found that machine learning methods that take site-specific conditions into account can improve prediction accuracy by a factor of two to three, depening on the error indicator chosen. Similar results are observed for multi-output ANNs for simulated in- and out-of-plane rotor blade tip deflection and root loads. Future work focuses on understanding transferability of results between different turbines within a wind farm and between different wind turbine types.

ONE DIMENSIONAL MODELLING FOR PULSED FLOW TWIN-ENTRY TURBINE

One-dimensional (1D) modelling is critical for turbomachinery unsteady performance prediction and system response assessment of internal combustion engines. This paper uses a novel 1D modelling (TURBODYNA) and proposes two additional features for the application to a twin-entry turbocharger turbine. Compared to single-entry turbines, twin-entry turbines enhance turbocharger transient response and reduce engine exhaust valve overlap periods. However, out-of-phase high frequency pulsating pressure waves lead to an unsteady mixing process from the two flows and pose great challenges to traditional 1D modelling. The present work resolves the mixing problem by directly solving mass, momentum and energy conservation equations during the mixing process instead of applying constant pressure assumption at the limb-rotor joint. Comparisons of TURBODYNA and an experimentally validated CFD suggest that TURBODYNA can not only provide a very good agreement on turbine performance, but also accurately capture unsteady features due to flow field inertial and pressure wave propagation. Levels of accuracy achieved by TURBODYNA have proved superior to traditional 1D modelling on turbine performance and the generality of the current 1D modelling has been explored by extending the application to another turbine featuring distinct characteristics.

Turbocharger Radial Turbine Response to Pulse Amplitude

Under on-engine operating conditions, a turbocharger turbine is subject to a pulsating flow and, consequently, experiences deviations from the performance measured in gas-stand flow conditions. Furthermore, due to the high exhaust gases temperatures, heat transfer further deteriorates the turbine performance. The complex interaction of the aerothermodynamic mechanisms occurring inside the hot-side, and consequently the turbine behavior, is largely affected by the shape of the pulse, which can be parameterized through three parameters: pulse amplitude, frequency, and temporal gradient. This paper investigates the hot-side system response to the pulse amplitude via Detached Eddy Simulations (DES) of a turbocharger radial turbine system including exhaust manifold. Firstly, the computational model is validated against experimental data obtained in gas-stand flow conditions. Then, two different mass flow pulses, characterized by a pulse amplitude difference of 5%, are compared. An exergy-based post-processing approach shows the beneficial effects of increasing the pulse amplitude. An improvement of the turbine power by 1.3%, despite the increment of the heat transfer and total internal irreversibilities by 5.8% and 3.4\%, respectively, is reported. As a result of the higher maximum speeds, internal losses by viscous friction are responsible for the growth of the total internal irreversibilities as pulse amplitude increases.

Multivariate Wind Turbine Power Curve Model Based on Data Clustering and Polynomial LASSO Regression

Wind turbine performance monitoring is a complex task because of the non-stationary operation conditions and because the power has a multivariate dependence on the ambient conditions and working parameters. This motivates the research about the use of SCADA data for constructing reliable models applicable in wind turbine performance monitoring. The present work is devoted to multivariate wind turbine power curves, which can be conceived of as multiple input, single output models. The output is the power of the target wind turbine, and the input variables are the wind speed and additional covariates, which in this work are the blade pitch and rotor speed. The objective of this study is to contribute to the formulation of multivariate wind turbine power curve models, which conjugate precision and simplicity and are therefore appropriate for industrial applications. The non-linearity of the relation between the input variables and the output was taken into account through the simplification of a polynomial LASSO regression: the advantages of this are that the input variables selection is performed automatically. The k-means algorithm was employed for automatic multi-dimensional data clustering, and a separate sub-model was formulated for each cluster, whose total number was selected by analyzing the silhouette score. The proposed method was tested on the SCADA data of an industrial Vestas V52 wind turbine. It resulted that the most appropriate number of clusters was three, which fairly resembles the main features of the wind turbine control. As expected, the importance of the different input variables varied with the cluster. The achieved model validation error metrics are the following: the mean absolute percentage error was in the order of 7.2%, and the average difference of mean percentage errors on random subsets of the target data set was of the order of 0.001%. This indicates that the proposed model, despite its simplicity, can be reliably employed for wind turbine power monitoring and for evaluating accumulated performance changes due to aging and/or optimization.

Horizontal Axis Wind Turbine Performance Analysis

The present work uses the method of Blade Element Momentum Theory as suggested by Hansen. The method applied to three blade models adopted from Rahgozar S. with the airfoil data used the data provided by Wood D. The wind turbine performance described in term of the thrust coefficient C_T, torque coefficient C_Q and the power coefficient C_p . These three coefficient can be deduced from the Momentum theory or from the Blade element Theory(BET). The present work found the performance coefficient derived from the Momentum theory tent to over estimate. It is suggested to used the BET formulation in presenting these three coefficients. In overall the Blade Element Momentum Theory follows the step by step as described by Hansen work well for these three blade models. However a little adjustment on the blade data is needed. To the case of two bladed horizontal axis wind