Predicting electricity consumption in the frame using a probabilistic neural network.

Abstract

The main objective of the study was to predict electricity consumption in the frame using a probabilistic neural network. Predicting electricity consumption using the PNN was successful since the performance of the model shows its capability as shown below; Goodness of fit: SSE: 1.96, R-square: 0.9214, Adjusted R-square: 0.9211, RMSE: 0.07494. Matlab programming was also successful as tool since it was able to perform its intended function during the prediction process. Basing on the success shown by PNN in predicting electricity consumption on a ring frame and because of the so many challenges faced using the other methods in predicting electricity consumption, I recommend the textile industries to employ the PNN model of prediction that will help in minimizing errors and save time. Also from the results, it shows that running low spinning speeds, using low efficiency motors and very high oil levels in the spindle bolster results in high electricity usage. Therefore, there is need always to optimize.