G. Mbamalu et al., DECOMPOSITION APPROACH TO FORECASTING ELECTRIC-POWER SYSTEM COMMERCIAL LOAD USING AN ARTIFICIAL NEURAL-NETWORK, Electric machines and power systems, 25(8), 1997, pp. 875-883
Different methods have been proposed and subsequently used in forecast
ing the power system load. Most of the reported work treat the system
load at the bulk power level. In practice, the system load is decompos
ed into user-based sectors. These are domestic, commercial, industrial
and municipal load sectors. Each sector is governed and influenced by
phenomena inherent to that sector. Thus the load consumption characte
ristic for a sector is unique due to different socio-economic function
s that take place in that sector. We use a multilayer neural network w
ith a back propagation algorithm to forecast the commercial sector loa
d portion resulting from decomposing the system load of the Nova Scoti
a Power Inc. system. To minimize the effect of weather on the forecast
of the commercial load, it is further decomposed into four autonomous
sections of six hour durations. The optimal input for a training set
is determined based on the sum of the squared residuals of the predict
ed loads. The input patterns are made up of the immediate past four or
eve hours load and the output is the fifth or the sixth hour load. ho
ur The results obtained using the proposed approach provide evidence t
hat in the absence of some influential variables such as temperature,
a careful selection of training patterns will enhance the performance
of the artificial neural network in predicting the power system load.