MODIFIED NARX NETWORK FOR LOW VOLTAGE CONSUMER LOAD PREDICTION

Authors

  • A. I. ABDULLATEEF Department of Electrical and Electronics Engineering, University of Ilorin
  • M. E. SALAMI Department of Electrical and Electronics Engineering, Elizade University
  • I. B. TIJANI Departmen of Electronics/Electrical, Abu Dhabi Men's College
  • M. T. AJALA Department of Industrial Maintenance Engineering, Yabatech. Lagos

Keywords:

Consumer load prediction, Data acquisition, Differential genetic, Algorithm evolution, Nonlinear autoregressive with eXogenous input.

Abstract

This paper is a continuation of our previous work on Nonlinear autoregressive with
eXogenous input (NARX) for load prediction. Application of NARX network in
real-time might be difficult since the tapped delay was selected by trial and error,
leading to nine hidden neurons which makes the network complex. This NARX
network was modified based on Genetic Algorithm (GA) and Differential Evolution
(DE) resulting in a new model coded NARX-DE-GA. GA and DE search for the
number of hidden neurons and tapped delay automatically. The NARX-DE-GA was
used to predict the consumer load using one month energy data with 8928 data
points. The results show that NARX- DE -GA outperformed the NARX network.
The training mean square error (MSE) value for NARX- DE -GA is 0.0253 while
validation is 0.0612. These values are slightly higher when compared with NARX
network in previous study which are 0.0225 and 0.0533 respectively. However, the
network structure which is one input and output tapped delay, and one hidden neuron
is simple and applicable in real time.

Downloads

Published

2019-09-25

Issue

Vol. 8 No. 2 (2019)

Section

Articles