The control of an induction motor (IM) drive is a complex process and requires speed estimation, which is dependent on various machine parameters. The linear regression approach reduces this dependency by eliminating the need for flux computation and gain adjustment of the PI controllers. The performance of the linear regression model deteriorates when data is noisy, the best fit deviates from the desired value. The Gaussian Process (GP) model is a non-parametric model, that can incorporate these noisy measurements and model uncertainties. In this paper, the speed of the IM drive with vector control is estimated using Gaussian Process Regression (GPR). The fictitious quantity$X$= ū*×$I$is used to eliminate the calculation of flux in the stator or rotor and its characteristic of stable drive in all four quadrants. The GP is viewed as a surrogate model, the prediction distribution obtained gives a confidence interval, used for usefulness validation that also distinguishes GP model from other black-box models. Furthermore, the Bayesian approach of GPR involves lesser complex evaluations making this approach simple and agile.

In transmission and distribution substations, power transformers account for the majority of capital investment. As they are expensive, effective thermal performance monitoring is necessary for life extension. The hot-spot temperature value is among the most significant factors affecting a transformer's life expectancy. As HST evaluation require some Top-oil Temperature model parameters, accurate estimations of TOT are required to analyze the thermal performance and lifespan of transformers. Conventionally, the TOT parameters from the Resistance-Capacitance network are determined using the input-output data. While the regressor signals meet the Persistence of Excitation criterion in the Gradient Estimator a parametric estimate error approaches zero and the parameter converges to their true value. The Design of Experiment is often carried out in the test center to meet the PE requirement. As a consequence, actual operating transformer data is utilized for the identification and estimation of the parameters associated with the TOT model by using finite-time estimators. The detected PE issue, the impact of DoE, and the efficiency of FTEs with different filters for non-PE data obtained from the actual operating transformers' thermal model are demonstrated by experimental analysis using MATLAB.

In the era of the vanishing of conventional energy sources, battery technology has gained tremendous demand. Thus, for the safe operation and the optimized utilization of the battery, it's modelling, and the state estimation is essential. However, in real-time situations, the accurate estimation of battery is quite hard and challenging because of its nonlinear characteristic and the influence of the various factors like driving load characteristics and operational conditions on battery performance. The state of charge (SOC) is a key parameter in the battery which gives an amount of the energy stored in the battery and it depends on various external factors like aging, temperature, and charging-discharging rate of the battery. In the literature, various methods exist for SOC estimation, however, these methods fail to consider the effect of external parameters. In view of this, the paper proposes a method in which the effect of the temperature is considered in SOC estimation by re-framing the existing state space battery model. This revised battery model is obtained by considering a temperature coefficient which illustrates the relation between SOC and the temperature. For the SOC estimation, an unscented Kalman filter (UKF) approach is used. Furthermore, the proposed method considers the influence of external factors as well as nonlinear characteristics of the battery. Finally, the proposed methodology is validated in MATLAB, for different temperature scenarios and the results shows the impact of temperature variation on SOC estimation.

Battery technology is an important component of an electric vehicle (EV). The modelling and state estimation of the battery are very relevant in the theoretical and practical operation. In addition to that, it helps to increase the battery life span, maximize its output, and reduces its cost. The state of charge (SOC) is a key parameter in the Battery Management System (BMS) because it helps to operate the battery in safe operating conditions. However, because of strong nonlinear characteristics of batteries, the internal states of the batteries cannot be measured directly. In the literature, various methods for SOC estimation have been developed. Though some of these methods are widely used in industries, some of these methods have their limitations like heavy computation cost, complex behaviour, etc. These methods fail to consider various parameters that affect the SOC of the battery. Thus, to address this issue the paper proposes a method which assists in the accurate estimation of SOC considering the parameters like temperature, aging, and charging-discharging rate of the battery. In this paper, a novel approach for the estimation of SOC of batteries with the modified state-space model using a Nonlinear Observer (NLO) is presented. The state equations derived from the second-order circuit model (Thevenin's model) is used to simulate a battery's complex dynamical behaviours. Further, the controller is designed using feedback linearization approach for optimal charging of the battery. The results obtained in MATLAB show a small error in estimated and actual SOC, along with this it highlights the effect of the temperature on SOC.

This paper presents a new approach to estimate the speed of vector-controlled three-phase induction motor drive using polynomial regression. The speed estimation via polynomial regression does not require any gain adjustment as well as PI controllers. The identification method is relying on the learning algorithm to enable an estimator fit between the reactive power (Q = Vsq.isd- Vsd.isq) and speed (ωr). The regressor quantity (Q) explicitly dependent on the voltage and current vectors and thus computation of flux is not required. The formulation of polynomial regression based speed estimation technique is simply realizable as the intensive computation of differential/integral is not required which makes the approach efficient and less time consuming. The Simulation results validates precise tracking of rotor speed for high speed, low speed and reversal from motoring to regenerative mode is achieved through this approach.