regression model

Predicting daily peak electricity consumption and identifying its determining factors are helpful in balancing electricity supply and demand. In this study, the maximum daily electricity consumption of Iran is modeled and predicted using different approaches in the period from March 20, 2017 to june 6, 2022. 90% of the observations are used to build the model and the rest are used to evaluate the predictive model. Based on the minimum prediction error criterion, the optimal neurons of the hidden layers of the neural network (NN) model were estimated to be 11 in the first layer and 8 in the second layer. The results show that the frequency of the first consumption peak at 11 a.m. and the frequency of the second consumption peak at 9 p.m. are higher than in the other hours. The variables holidays, hours of energy consumption, number of new corona patients, temperature and relative humidity of the air and population have a significant influence on peak electricity consumption in the regression model. The influence of holidays and air temperature is also stronger than other variables. Comparing the peak power consumption prediction results shows that the prediction accuracy of different models and approaches is not the same. The average prediction error percentage of the GLM, NN, and ARIMA models during 187 days (02/12/2021 - 06/06/2022) are 0.0799, 0.0754, and 0.0714, respectively. Therefore, the ARIMA model with the minimum average forecast error is a suitable model for peak consumption forecasting.

In this study, mechanical alloying process for producing of aluminum-based nanocomposites reinforced by SiC nanometric particles was modeled by using statistical methods. First, more effective parameters have been selected for analyzing. Since, in order to obtain nanocomposit powder, practical experiments have been planned and carried out. Because, the mechanical alloying process is affected by many complex parameters, complete physical modeling of the process seems to be almost impossible in current condition. Therefore, in this study, statistical modeling was used to analyze the process. By performing statistical analysis of the obtained parameters from experiments, variation trend and mean particle size were perceived. Then, using the regression method, the desired mechanical alloying process was modeled by considering the selected input and output parameters. Statistical analysis was performed in order to make sure about the suitability of this model and it has become clear that the mean particle size is affected by ball to powder weight ratio, squared the ball to powder weight ratio, and the interaction of multiplication of milling speed by milling time. Other input variables in regression model were of little importance and the effect of these terms in the final regression model was neglected.