Estimation of electric power load on electric power substation is an essential task for system operator in order to operate the system in a reliable and optimal manner. In this paper, machine learning with artificial neural network is used for forecasting the load at a particular hour of the day on an electric power substation. Historical load data at each hour of the day for the period from September-2018 to November-2018 is taken from 33/11 kV substation near Kakatiya University in Warangal. A new artificial neural network architecture is developed based on the approach used to forecast the load. The developed model is simulated in MATLAB with available historical data to forecast the load on 33/11 kV electric power substation. Based on the analysis it is observed that the proposed architecture forecasts the load with better accuracy.Keywords Artificial neural networks · Electric power load forecasting · Machine learning · Mean square error · Mean absolute percentage error List of symbols L(D, t) Load at Dth day and tth hour L(D, t − 1) Load at Dth day and (t − 1)th hour L(D, t − 2) Load at Dth day and (t − 2)th hour L(D, t − 3) Load at Dth day and (t − 3)th hour L(t, D − 1) Load at (D − 1)th day and tth hour L(t, D − 2) Load at (D − 2)th day and tth hour L(t, D − 3) Load at (D − 3)th day and tth hour L(t, D − 4) Load at (D − 4)th day and tth hour MAPE Mean absolute percentage error MSE Mean square error y Target i Actual output y Predicted i Predicted output m Number of samples R Regression coefficient * Venkataramana Veeramsetty,

It is already established that the renewable integration effects to the power system are nonzero and become more important with large penetrations. Thereby, the impacts of renewable energy sources (RESs) after integration are studied in this work to stabilize grid frequency of the studied test power system model. Initially, the two-area power system model is studied as the test system. The purpose is to show the tuning efficiency of non-conventional quasi-oppositional dragonfly algorithm (QODA) algorithm as compared to conventional way of tuning technique. It is showed that QODA algorithm is quite effective to find the optimal parameters of proportional-integral-derivative (PID) controller in load frequency control performance. Further, the three-area power system model integrated with RESs is studied. The work done here is to study the impacts of wind turbine generation, solar thermal power generation and solar photovoltaic on system frequency oscillations. The PID controller is employed as the supplementary control task, and its parameters are tuned by QODA algorithm. The integral of time absolute error is chosen as the objective function, and further performance indices are determined at the end of the execution of the program to examine the performance of the designed QODA-based PID controller. Following the integration of RESs, the impacts on frequency deviation through simulation results are also presented. The simulation results showed that the RESs are quite effective in regulating the power system frequency deviation understudied.

A pulse-width modulation (PWM) inverter and sinusoidal output voltage supplies feeding four different chorded three-phase induction motors were tested for low-order odd voltage harmonic components and efficiency at different loads. The total harmonic distortion due to the third, fifth, and ninth harmonics was lowest in a motor with 160° coil pitch energized by both sinusoidal and PWM voltages. The efficiencies of the motor with the short-chorded winding were as much as 5% and 16% higher than that of the full-pitched motor under sinusoidal and PWM excitation, respectively, due to harmonic cancellation.

No abstract

No abstract