t3 load forecasting
TRANSCRIPT
-
8/12/2019 T3 Load Forecasting
1/3
Load Forecasting
Load demand forecasting is essentially important for the electric companies. It has many
applications including energy purchasing and generation planning, load switching, contract
evaluation, and infrastructure development. A large variety of mathematical methods have been
developed for load forecasting. Many variables and factors could affect the load forecasting. The
most important factor is load peak for the given regions for at least for the last 5 years.
Forecasting refers to predict how greatly power will be expected to supply where and when the
power must be delivered. Electric load forecasting can be divided into three categories, namely:
Short-term load forecasting, Medium-term load forecasting and, Long-term load forecasting.
The short-term load forecasting predicts load demand in time interval from one day to several
weeks. It can be used to approximate load flows to make decision that can avoid overloading.
Therefore, this will help in developing network reliability and decreases occurrences of equipmentfailures and blackouts.
The medium-term load forecasting predicts the load demand from a month to several years. This
will provides essential information for future planning operation.
The long-term load forecasting predicts the load demand in time interval from a year up to twenty
years, it is essentially for power system planning, utility development, and employees hiring.
Forecasting Techniques
Forecasting techniques relays only on the extrapolation of past observation of the load using
mathematical procedure to extrapolate them into the future. One of the forecasting techniques is the
time-series models. These models depend on the fitting of a time series to original data. The load,
PL(t), is expressed as a function of time t,f(t). Some of the models are:
Straight Line PL(t) =a + b.t
Time Polynomial PL(t) =i
i
itb
Exponential PL= a.ebt
The parameters of the above models can be estimated using the least-square technique.
Method of Least Squares
Example 1- Linear Model
Based on the annual peak load data given in Table 1
a. Find the function of the form
y= c1+ c2x
that is the best least-squares fit to the data points.
b. Hence estimate the annual peak load and growth for five more years (6-10).
1
-
8/12/2019 T3 Load Forecasting
2/3
Table 1 Substation annual peak loads
Solution:
The least squares solution that minimizes the least-square error is given by
( )
.4520
17920.
1
1
1
....
2
1
2
12
1
2
1
1
=
=
=== +
C
C
C
C
x
x
x
y
y
y
YXXXYXCCXY
nn
TT
MMM
kWxy 452017920+=
Based on this model the annual peak load for the subsequent five (6-10) years can be estimated. The
results are shown in Table 2
Table 2: Estimated annual peak loads for years 6 to 10
kWxy 452017920+=
y = 4520x + 17920
R2= 0.9377
0
10000
20000
30000
40000
50000
60000
70000
0 2 4 6 8 10Year
AnnualPeakLoad,k
12
Given Data Forecas ted Linear (Given Data)
2
-
8/12/2019 T3 Load Forecasting
3/3
Example 2- Exponential Model
Based on the annual peak load data given in Table 1
a. Find the function of the form
y=c1exp(c
2x) kW
that is the best least-squares fit to the data points.
b. Hence estimate the annual peak load and growth for five more years (6-10).
Based on this model the annual peak load for the subsequent five (6-10) years can be estimated. The
results are shown in Table 3
Table 3 Estimated annual peak loads for years 6 to 10
Year Peal Load, kW
6 48174
7 55947
8 64976
9 75461
10 87638
y=19633*exp(0.1496*t) kW
y = 19633e0.1496x
R2= 0.9355
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
0 2 4 6 8 10
Year
AnnualPeakLoad,k
12
Given Data Forecasted Expon. (Given Data)
3