The increasing integration of solar photovoltaic (PV) systems into modern energy grids presents significant challenges due to the intermittent and weather-dependent nature of solar energy generation. Accurate short-term forecasting is essential to ensure grid stability and optimize energy resource allocation. This study proposes a comprehensive data-driven framework for solar energy forecasting using multiple machine learning (ML) techniques, including Multiple Linear Regression, Ridge, Lasso, Decision Tree Regression, Support Vector Regression, and ensemble-based models such as Random Forest, AdaBoost, Bagging, and Gradient Boosting Regressors. The framework incorporates advanced feature engineering using high-resolution meteorological and solar geometric parameters-such as relative humidity, temperature, cloud cover, zenith angle, azimuth, and angle of incidence-to enhance model accuracy. Historical solar power and weather datasets were used to train and evaluate the models across multiple performance metrics. Among the models, the Gradient Boosting Regressor demonstrated the best performance, achieving an R2 of 0.827, RMSE of 399.44, and MAE of 253.62, marking a significant improvement over baseline models. The study also evaluates model robustness and discusses feature relevance, hyperparameter optimization strategies, and deployment considerations for real-time grid operations. These findings provide practical insights for stakeholders aiming to implement intelligent solar forecasting systems in smart grid environments, thereby contributing to enhanced energy management and grid resilience.
A novel neuro‐adaptive control scheme is proposed in the context of angular velocity tracking in DC–DC buck converter driven permanent magnet DC motor system. The controller builds upon the idea of backstepping and consists of a fast single hidden layer Hermite neural network (HNN) module equipped with on‐board (adaptive) learning to counteract the unknown non‐linear time‐varying load torque and to ensure nominal tracking performance. The HNN has a simple structure and exhibits promising speed and accuracy in estimating dynamic variations in the unknown load torque apart from being computationally efficient. The proposed method guarantees a rapid recovery of nominal angular velocity tracking under parametric and non‐parametric uncertainties. In order to verify the performance of the proposed neuro‐adaptive speed controller, extensive experimentation has been conducted in the laboratory under various real‐time scenarios. Results are obtained for start‐up, time‐varying angular velocity tracking and under the influence of highly non‐linear unknown load torque. The performance metrics such as peak undershoot/overshoot and settling time are computed to quantify the transient response behaviour. The results clearly substantiate theoretical propositions and demonstrate an enhanced dynamic speed tracking under a wide operating regime, thus confirming the suitability of proposed method for fast industrial applications.
To use a Hybrid Excitation Synchronous Machine (HESM) in a hybrid electrical vehicle (HEV), its performance indicators such as back-EMF, inductance and unbalanced magnetic force should be computed preferably by an analytical method. First, the back-EMF is calculated by considering alternate-teeth and all-teeth non-overlapping and overlapping windings. The effects of three types of magnetization patterns including the radial, parallel and Halbach magnetizations on the back-EMF waveform have also been investigated. Then, the self-inductance of the stator and rotor windings, the mutual inductance between the stator and rotor windings, and the mutual inductance between the stator phases are computed. Next, the components of the unbalanced magnetic force (UMF) in the direction of the x and y axes and its amplitude are computed. Moreover, the effects of the magnetization patterns on those magnetic pulls are investigated. To minimize the UMFs, symmetry must be implemented in the excitation sources; therefore, first the stator winding then the permanent magnet and rotor winding are modified in such a way that the UMFs are reduced. Increasing the temperature leads to a weakening of the magnet’s residual flux density, which strongly affects the performance characteristics of the electric machine such as Back-EMF and UMF. Finally, the ratio of the permanent magnet flux to the rotor flux is determined in such a way that the average torque is maximized. In this section, the effects of three magnetization patterns will be investigated.
2 May 2022 in International Journal of Fuzzy Systems
Direct torque control (DTC) is one of the most prominent control techniques used by permanent magnet synchronous motor (PMSM) drives in industry applications. Nevertheless, the presence of hysteresis controllers and inaccurate voltage switching table in traditional DTC results in higher torque and flux ripple. This study proposes an enhanced DTC functioned Surface‐mounted PMSM (S‐PMSM) drive with mitigation of torque and flux ripple. The operation relies on generating the reference voltage vector (VV) in a stationary reference frame, which supports control of torque and flux without hysteresis controllers. The reference VV generation is simple and does not affect control robustness. The position of reference VV in a sector is used to build the voltage vector (VV) switching table. As a result, the application of nearest discrete VV to reference VV produces optimal torque and flux control. Moreover, redundant switching combinations of null VV are effectively used for possible minimization of switching frequency of two‐level voltage source inverter (VSI) supplied S‐PMSM drive. Therefore, proposed DTC gains improved S‐PMSM drive response along with switching frequency reduction. In dSPACE‐RTI 1104 platform, experimental response of S‐PMSM drive under various operating conditions have been depicted to highlight the proficiency of proposed DTC in comparison with existing DTC.