Short Term Load Forecasting Using Artificial Neural Network

Shivam A. Koli

Department of Electrical Engineering, walchand college of engineering, Vishrambag sangli, Maharashtra

416415, India

Abstract.
Short-term electricity demand forecasting is crucial for the effective control and scheduling of power
systems. Traditional econometric approaches, which often assume linear relationships between weather
variables and load demand, may not provide accurate results due to the nonlinear and non-stationary nature
of these relationships. This paper proposes an adaptable technique using Artificial Neural Network (ANN)
for load forecasting. ANNs, utilize supervised learning to map historical data and various inputs to forecast
future loads. The proposed model employs a single-layer neural network with one input layer, one hidden
layer, and one output layer. Key input variables include hour and day indicators and weather-related inputs.
The ANN’s ability to adapt and learn from data makes it a valuable tool for optimizing unit commitment,
startup and shutdown of power plants, control of spinning reserves, and power trading in interconnected
systems.

Keywords: load forecasting; artificial neural network.

1.Introduction- commitment, start up and shut down of thermal

plants, control of spinning reserve and buying and
Short-term forecasts of electricity demand selling of power in interconnected systems.an
are needed for control and scheduling of power advantage of the Artificial neural network is that it
systems, and hence they are very important for does not require any assumption between load and
dispatchers. Short-term electric load demand is weather variables in advance. Moreover, the
often considered as a function of weather variables literature suggests that the accuracy of forecasts is
and human social activities. Traditional better than traditional econometric forecasting.
econometric approaches establish functional In this literature we propose a method of
relationships between weather variables and current load forecasting based on a single-layer neural
load demand for the forecasting function, often network and input variables to be used for such
assuming a linear relationship. However, as Park et neural network.
al. (1991) indicate, the econometric approach may
not give sufficiently accurate results because of 2.Artificial neural network
nonlinear and non-stationary relationships between
the load data and weather variables. Therefore, an Artificial neural network is an information
adaptable technique is needed. paradigm that is inspired by biological nervous
An artificial neural network (ANN) systems. The system structure is composed of a
approach is such an adaptable technique. The large number of highly interconnected processing
Artificial neural network uses a supervised learning elements (neurons) working together to solve
technique for training neural networks. The goal of specific problems. Each unit operates only on its
this type of network is to create a model that local data and on the inputs, they receive via the
correctly maps the input to the output using connections. The proposed neural network consists
historical data and various inputs so that the model of one input layer, one hidden layer and one output
can then be used to produce the forecast. These layer. The neural network maps the input vector X
load forecasts play a vital role in optimum unit to the output vector Ô (Fig. 1).
 Figure 2: hourly Load curve in
India
3.2. Day indicator D(n)
Figure 3:neural network structure.
Load also changes throughout the week it
The output is a sum of the weighted nonlinear is clear from the fig.3 that power consumption in
terms given by where the weekend (day 6,7) drops compared to power
consumption in the weekdays. To take this into
account while calculating load forecasting, variable
associated with day of week D(n) where n changes
from 1 to 7 representing days of week
Wᵢⱼ (i = l,..., n ; j = l ,..., nᵢ) is the weight factor
between the ith input node and the ith hidden node;
v ᵢ is the threshold value at the ith hidden node
defined by the algorithm; and xⱼ is the jth input
element.
The learning of a network is a recursive
process which adapts the connection strengths, i.e.
the weights among the nodes and the threshold
values at each iteration. The error function E is to
update the weights at each recursive minimum:

Figure 1: load changes during

where K is the number of cases used to train the week
network O(m) is the actual output of the training
case m, and Ô(m) is the forecasted output of the
network.

3.selection of input variables

The most important work in building an
Artificial neural network load forecasting model is
the selection of input variables. Selection of inputs
is based on judgement of experience of the
engineer and it is carried out by trial and error since
there is no rule of thumb that can be followed to
determine which inputs to use. However, some
statistical analysis can be done to determine the
variables with most influence on power
consumption. These inputs are (a) hour and day, (b)
weather associated inputs and (c) historical loads
3.1. Hour indicator H(i)
Load changes throughout the day from one hour to
another. On a typical day in India peak load occurs Figure 3: load changes during
generally between 06:00PM to 08:00PM. And week
minimum load occurs at 03:00AM(Fig.2). We can
conclude that the magnitude differs between
various hours by a large margin there for an hour
indicator H(i) (where i changes between 1-24) is
very helpful while calculating short term load Figure 4: monthly power consumption and
temperature graph
3.3. Weather indicating variables optimization algorithms, such as gradient
descent or stochastic gradient descent.
Weather variables primarily include
temperature, wind velocity cloud coverage, rain 5. Performance measures of the network
etc. out of all the above variables temperature
seems to have strongest correlation to power In order to evaluate the performance of the
consumption as displayed by Fig.4. Although wind network model, the following common
velocity and cloud coverage have their own effects performance measures are used automatically by
on load due to its effect being very small no the software: mean square error (MSE), normalized
significant relation is seen. therefore, only mean square error (NMSE), correlation coefficient,
temperature as a weather variable is considered % error, and the minimum description length
while calculating load forecasting (MDL). The most common measure is to minimize
MSE in the forecasting but this measure has some
3.5. Normalized data drawbacks. Alternative measures such as
percentage error, mean percentage error (MPE) and
Normalization or scaling is not really a mean absolute percentage error (MAPE) are also
functional requirement for the neural networks to presented in this study
learn, but it significantly helps as it transposes the
input variables into the data range that the sigmoid 6.limitations of ANN
activation functions lie in (i.e. for logistic [0, 1] and
tanh [-1, 1]. And it makes it so that all inputs are in A major problem in modelling load
comparable range forecasting is that there are usually many trends in
the data. For example, there may be a seasonal
Each input data is multiplied by an trend reflecting the seasons of the year. As well as
amplitude and shifted by an offset. The amplitude there may be monthly and/or weekly trends in the
and offset are often referred to as normalization data as well. These trends make it difficult to
coefficients. The normalization coefficients are correctly model load forecasting data.
computed based on the minimum and maximum
values found across all of the data sets. The Artificial neural networks can become too
formulae used to compute these coefficients are complex because of their architecture and the huge
datasets used to train them. They can also
memorize the training data. These can lead to poor
generalization of new data.

where Max(i) and Min(i) are the maximum and They can also be vulnerable to adversarial
minimum values The input components normalize examples or minor changes in the input data. These
changes can cause the particular artificial neural
network to make incorrect decisions and irrelevant
outcomes.
the data using the following formula:

4.Training of ANN 7. Conclusion

The application of Artificial Neural
Training is a process to find the input–
Networks (ANNs) for short-term electricity load
output map by repeatedly analyzing the training set
forecasting presents a significant advancement over
as the network learns. The parameters of the
traditional econometric methods. By leveraging the
network obtain optimal values during this process
adaptability and learning capabilities of ANNs, this
and the performance checked with the test set. The
approach addresses the nonlinear and non-
ingredients for supervised learning are the input,
stationary relationships between load demand and
the desired response, the definition of error and a
influencing variables such as weather and human
learning law.
activities. The proposed single-layer neural
Training can be implemented in two network model, which incorporates key input
ways. Either we present a pattern and adapt the variables like hour and day indicators and weather-
weights (on-line training), or we present all the related inputs demonstrates improved forecasting
patterns in the input file (an epoch), accumulate the accuracy.
weight updates, then update the weights with the
This enhanced accuracy is critical for
average weight update, called batch learning.in
optimizing power system operations, including unit
proposed ANN backpropagation can be
commitment, thermal plant management, spinning
implemented to effectively train the artificial neural
reserve control, and power trading. The findings
network. Backpropagation is an iterative algorithm,
underscore the potential of artificial neural
that helps to minimize the cost function by
networks to enhance load forecasting, offering a
determining which weights and biases should be
efficient and flexible solution that can adapt to the
adjusted. During every epoch, the model learns by
dynamic nature of electricity demand.
adapting the weights and biases to minimize the
loss by moving down toward the gradient of the Bibliography
error. Thus, it involves the two most popular
(1) Ongsakul, Weerakorn_ Vo, Dieu Ngoc -
Artificial intelligence in power system optimization
(2013, CRC Press, Taylor & Francis)
(2) P.K. Dash et al./Electric Power Systems
Research 32 (1995) 19-23
(3) O.A.S. Carpinteiro et al. / Applied Soft
Computing 4 (2004) 405–412
4) N. Kandil et al. / Electrical Power and Energy
Systems 28 (2006) 525–530
(5) S. C. Bhattacharya and L. T. Tha Short - Term
ELECTRIC LOAD forecasting
(6) S.M. Islam et al./Electric Power Systems
Research 34 (1995) 1-9
(7) M. Ghiassi et al. / Electric Power Systems
Research 76 (2006) 302–316