03 - Lectures of Load Forecasting - Nnat

SHORT TERM LOAD FORECASTING
USING NEURAL NETWORKS AND

FUZZY LOGIC
Arizona State University
06/2000 Short-term load forecasting 1

Short Term Load Forecasting
Content
Overview of Short term load forecasting
Introduction
Definitions and expected results
Importance of Short-term load forecasting
Impute data and system parameters required for load forecasting
Concept
Major load forecasting techniques
Concept of STLF model Development
Statistical methods
(Multiply liner regression,
stochastic time series e.t.c)

Short Term Load Forecasting
Content
Artificial Neural Networks
Building block: a feed forward network
Load forecasting engine
Results
Numerical example
Fuzzy logic and evolutionary programming

Overview of Short Term
Load Forecasting (STLF)

Introduction
The electrical load increases about 3-7%
per year for many years.
The long term load increase depends on
the population growth, local area
development, industrial expansion e.t.c.
The short term load variation depends on
weather, local events, type of day
(Weekday or Holiday or Weekend) e.t.c.
Introduction
The building of a power plant requires:
10 years (Nuclear)
6 years (Large coal-fired)
3 years (combustion turbine)
The electric system planing needs the
forecast of the load for several years.
Typically the long term forecast covers a
period of 20 years
Introduction
The planning of maintenance, scheduling
of the fuel supply etc. calls for medium
term load forecast .
The medium term load forecast covers a
period of a few weeks.
It provides the peak load and the daily
energy requirement

Introduction
The number of generators in operation,
the start up of a new unit depends on the
load.
The day to day operation of the system
requires accurate short term load
forecasting.

Introduction
Typically the short term load forecast
covers a period of one week
The forecast calculates the estimated load
for each hours of the day (MW).
The daily peak load. (MW)
The daily or weekly energy generation.
(MWh)
Introduction
The utilities use three types of load
forecasting:
Long term (e.g. 20 years)
Medium term. (e.g. 3-8 weeks)
Short term (e.g. one week)
This lecture presents the short term load

forecasting techniques
Definitions and Expected Results
(Big picture)
The short term load forecasting provides

load data for each hour and cover a
period of one week.
The load data are:
hourly or half-hourly peak load in kW
hourly or half-hourly values of system
energy in kWh
Daily and weekly system energy in kWh
(Big picture)
The short term load forecasting is

performed daily or weekly.
The forecasted data are continuously
updated.
Typical short-term, daily load forecast is
presented in the Table in the next page.
(Salt River Project, SRP)

(Big picture)
Typical short-term, daily load forecast from Salt

River Project. Hourly Load in MW
1/01/85
2153 2075 2001
2787 2677 2545
01/02/85
1930 1893 1893
2721 2613 2500
01/03/85
1892 1839 1835
2657 2526 2412
01/04/85
1856 1803 1806
2716 2620 2537
01/05/85
1793 1720 1672
2509 2314 2170
(Big picture)
Typical short-term, daily load forecast from Salt

River Project. Hourly Load in MW
1/01/85
2153 2075 2001
2787 2677 2545
01/02/85
1930 1893 1893
2721 2613 2500
01/03/85
1892 1839 1835
2657 2526 2412
01/04/85
1856 1803 1806
2716 2620 2537
01/05/85
1793 1720 1672
2509 2314 2170

Importance of Short-term Load
Forecasting
Provide load data to the dispatchers for

economic and reliable operation of the
local power system.
The timeliness and accuracy of the data
affects the cost of operation.
Example: The increase of accuracy of the
forecast by 1% reduced the operating cost
by L 10M in the British Power system in
1985
Forecasting
The forecasted data are used for:

Unit commitment.
selection of generators in operation,
start up/shut down of generation to minimize
operation cost
Hydro scheduling to optimize water release
from reservoirs
Hydro-thermal coordination to determine the
least cost operation mode (optimum mix)
Forecasting
The forecasted data are used for:

Interchange scheduling and energy purchase.
Transmission line loading
Power system security assessment.
Load-flow
transient stability studies
using different contingencies and the predicted
loads.
Forecasting
These off-line network studies:

detect conditions under which the system is
vulnerable
permit preparation of corrective actions
load shedding,
power purchase,
starting up of peak units,
switching off interconnections, forming islands,
increase spinning and stand by reserve
Impute Data and System
Parameters for Load Forecasting
The system load is the sum of individual

load.
The usage of electricity by individuals is
unpredictable and varies randomly.
The system load has two components:
Base component
Randomly variable component

The factors affecting the load are:

economical or environmental
time
weather
Unforeseeable random events

Economical or environmental factors

Service area demographics (rural, residential)
Industrial growth.
Emergence of new industry, change of farming
Penetration or saturation of appliance usage
Economical trends (recession or expansion)
Change of the price of electricity
Demand side load management
The time constraints of economical or

environmental factors are slow,
measured in years.
This factors explains the regional variation
of the load model (New York vs. Kansas)
The load model depends on these slow
changing factors and has to be updated
periodically
Time Factors affecting the load

Seasonal variation of load (summer,
winter etc.). The load change is due to:
Change of number of daylight hours
Gradual change of average temperature
Start of school year, vacation
Calls for a different model for each

season
Typical Seasonal Variation of Load

Summer peaking utility


Daily variation of load. ( night, morning,etc)

Weekly Cyclic Variation

Saturday and Sunday significant load reduction
Monday and Friday slight load reduction
Typical weekly load pattern:


Holidays (Christmas, New Years)
Significant reduction of load
Days proceeding or following the holidays
also have a reduced load.
Pattern change due to the tendency of prolonging
the vacation

Weather factors affecting the load

The weather affects the load because of
weather sensitive loads:
air-conditioning
house heating
irrigation

Weather factors affecting the load

The most important parameters are:
Forecasted temperature
Forecasted maximum daily temperature
Past temperature
Regional temperature in regions with
diverse climate

Weather Factors Affecting the Load

The most important parameters are:
Humidity
Thunderstorms
Wind speed
Rain, fog, snow
Cloud cover or sunshine

Random Disturbances Effects on Load

Start or stop of large loads (steel mill,
factory, furnace)
Widespread strikes
Sporting events (football games)
Popular television shows
Shut-down of industrial facility
The different load forecasting techniques

use different sets of data listed before.
Two -three years of data is required for
the validation and development of a new
forecasting program.
The practical use of a forecasting
program requires a moving time window
of data
The moving time window of data

requires:
Data covering the last 3-6 weeks
Data forecasted for the forecasting period,
generally one week

The selection of long periods of historical

data eliminates the seasonal variation
The selection of short periods of historical

data eliminates the processes that are no
longer operative.

The forecasting is a continuous process.

The utility forecasts the load of its service
area.
The forecaster
prepares a new forecast for everyday and
updates the existing forecast daily
The data base is a moving window of data
Major Load Forecasting
Techniques
Statisticalmethods
Fuzzy logic
Evolutionary programming
Simulated Annealing and expert system
Combination of the above methods

Major Load Forecasting
Techniques
The statistical methods will be discussed

briefly to explain the basic concept of the
load forecasting
This lecture concentrates on load

forecasting methods using neural
networks and fuzzy logic

Concept of STLF Model
Development
Model selection
Calculation and update of model
parameters
Testing the model performance
Update/modification of the model if the
performance is not satisfactory

Development
Model selection
Selection of mathematical techniques that
match with the local requirements
Calculation and update of model
parameters
This includes the determination of the constants and
selection of the method to update the constants
values as the circumstance varies. (seasonal
changes)
Development
Testing the model performance

First the model performance has to be
validated using 2-3 years of historical data
The final validation is the use of the model in
real life conditions. The evaluation terms
are:
accuracy
ease of use
bad/anomalous data detection
Development
Update/modification of the model if the

performance is not satisfactory
Due to the changing circumstances (regional
gross, decline of local industry etc.) the
model becomes obsolete and inaccurate,
Model performance, accuracy has to be
evaluated continuously
Periodic update of parameters or the change
of model structure is needed
for STLF

for STLF
Several Artificial Neural Network (ANN)

based load forecasting programs have
been developed.
The following neural networks were
tested for load forecasting:
Feed-forward type ANN
Radial based ANN
Recurrent type ANN
Building Blocks of
a Feed Forward Network
A Feed forward Three-Layered

Perceptron Type ANN was selected to
demonstrate the short term load
forecasting technique.
The selected network forecasts:
Hourly loads
Peak load of the day
Total load of the day.

Building Blocks of
The forecasting with neural network will

be demonstrated using a feed forward
three-layered network to forecast the
peak load of the day.
The network has:
one output: (load in kW)
three input: (previous day max. load and
temperature, forecasted max. temperature)_
Building Blocks of
The structure of the Feed Forward

Three-Layered Perceptron Type ANN is
presented on the next page.
The network contains:
i = 1.. 3 input layer nodes
j = 15 hidden layer nodes
k=1 output layer nodes

for STLF
Y
w1 w2 w3 w4 w5
1 2 3 4 5
W11 W12 W25 W35
1 2 3
X1 X2 X3

Building Blocks of
The inputs are:

X1 previous day max. load
X2 previous day max. temperature
X3 forecasted max. temperature
Wij weight factor between input
and hidden layer
wj weight factor between hidden
layer and output
Building Blocks of
A sigmoid function is placed in the nodes

(neurons) of the hidden layer and output
node.
The sigmoid equation for an arbitrary Z
function is:
1
1 e Z
Y output :maximum load

Building Blocks of
Inputs Xi are multiplied by the

connection weights (Wij) and passed on to
the neurons in the hidden layer.
The weighted inputs (Xi*Wij) to each
neurons are added together and passed
through a sigmoid function.
Input of hidden layer neuron 1 is:
X1 W11 X2 W21 X3 W31

Building Blocks of
The output Hj of the jth hidden layer

node is:
Hj 1
n
W X
ij j
1 e i 1
Building Blocks of
Inputs Hj are multiplied by the

connection weights (wk) and passed on to
the neurons in the output layer.
The weighted inputs (Hj*wk) to each
output neurons are added together and
passed through a sigmoid function.
Input of output neuron is:
H1 W1 H2 W2 H3 W3 H4 W4 H5 W5

Building Blocks of
The output Y is:
Y 1
h
w H
j 1 j j
1 e
Training of the
Feed Forward Neural Network
The described neural network is trained

using historical data.
Typical data set contains 2-3 years of
load and weather data.
Error back propagation (BP) method is
used for the training.
During the learning the weights are
adjusted repeatedly.
Training of the
The output produced by the ANN in

response to inputs are repeatedly
compared with the correct answers
Each time the weights are adjusted
slightly by beck-propagating the error at
the output layer through the ANN
Equations for the training are presented
in the next page
Training of the
The equations used for the training are:

Weight is between input and hidden layer:
n 1
Wij Y ( Yactual Y ) (1 Y ) H j
n
Wij
w j (1 H j ) X j
Weight is between hidden layer and output:

w nj1 w nj Y ( Yactual Y ) (1 Y ) H j

Training of the
In the equations:
Yactual is the true value of the output load
is learning factor (0.3-0.8)
n is the number of learning cycles
Xj is the input value belongs to Yactual
A numerical example demonstrates the
use of neural forecasting method.
Training of the
The over training has to be avoided using

the cross validation method:
The training set is divided into two parts.
(Part 1 : two years data, Part 2: one year
data)
Part 1 is used to train the network, by
passing the data through the network.
Few hundred times pass represents a
training period.

Training of the
The over training of the network has to

be avoided:
Part 2 is used to check the effectiveness of
the training.
After each training period the error is
calculated when the network is supplied by
the input data of Part 2.
The increase of error indicates over training,
when the training has to be stopped
Load Forecasting Engine
The EPRI developed a Load Forecasting

Engine using 24 Neural networks.
One network forecasts the load for each
hour of the day.
The networks are grouped into four (4)
categories depending on time of the day.
The categories have different inputs.

The construction of the engine shows the four

groups of neural networks.

The four categories are:

Category 1. Nine neural networks. Forecasts
the load between 1-9AM .
Category 2. Nine neural networks. Forecasts
the load between 10AM -2PM and 7 -10PM .
Category 3. Four neural networks. Forecasts
the load between 3-6 PM .
Category 4. Two neural networks. Forecasts
the load between 11-12 PM .
The inputs in four categories are:

Category 1. Forecast for early morning
general input
load in the last three-four hours
temperature in the last three-four hours
Category 2. Forecast for off peak hours
general input
forecast temperature of previous hours
yesterdays load and temperature of hours close
to this hours
The inputs in four categories are:

Category 3. Forecast for afternoon peak hours
general input
forecast temperatures of previous and feature
hours close to this hours.
yesterdays load and temperature of hours close
to this hours
Category 4. Forecast for late night hours
general input
forecast temperatures for the four proceeding
hours
The general input variables are

same hour load, temperature and humidity
of one day ago. (3)
same hour load, temperature and humidity
of two days ago. (3)
same hour load and temperature seven (7)
days ago (2)

The general input variables are

(continuation):
same hour forecast temperature and relative
humidity of the next day (2)
day of the week index (Sunday 01, Monday
02 etc.)

The load forecasting engine has one

output for the hourly load
The extended forecast uses the forecasted
values. E.g. The two-day ahead forecast
uses values obtained by the one-day
ahead forecast.
The forecast can be updated each hour
using the recent load and weather data
The weights in the neural network are

adjusted daily.
The retraining uses the actual load and
weather data of the past few days.
The retraining helps to follow the trends,
changes of weather patter e.t.c

Load Forecasting For Holidays
The load during the holidays has
different patterns and is significantly
reduced.
The forecast is inaccurate because of the
small number of historical data.
The holiday is treated as
Saturday if the shopping centers are open
Sunday if the shopping centers are not open
Hourly Weather Forecast
The weather service provides forecasts

for:
daily maximum and minimum temperature
daily maximum and minimum relative
humidity
rain and fog
maximum wind speed
No hourly data are provided.
Hourly Weather Forecast
EPRI developed an hourly temperature

and humidity forecasting engine.
Single neural network with
28 inputs :
hourly temperature of the previous day)
high and low temperature of the two previous
day
24 outputs: expected hourly temperatures.
Load Forecasting Results

Comparison of forecasted and actual loads

Accuracy less than 3% for the next days forecast is

considered good
The longer term forecast accuracy is less (7-8%)

Appendix 1
Derivation of Learning
Algorithm

Algorithm
The output of the hidden and output layer and the error
function are:
1 1
H Y
j n h
W X w H
ij j j j
i1 j 1
1 e 1 e
1
E ( X 2 ,n Y ) 2
2

Algorithm
The update of the weight factors require iteration
dE dE
dw j dWi j
dw j dWi j
For the calculation of the derivative the following

substitutions are used
h
Y1 w H Y 2 1 e Y1 u ( X 2 ,n Y )
j j
j 1
n
h1 W X h 2 1 e h1
ij i
i1
Algorithm
After substitutions the equations are:
1 1 1
H
j I h1 h2
W X 1 e
ij i
i1
1 e
1 1 1
Y
J 1 e Y1 Y 2
w H
j j
j 1
1 e
1 1
E ( X 2 ,n Y ) 2 u 2
2 2
Algorithm
The derivation of the error function results in :
dE d 1 2 du du du
2 u dw u dw ( X 2 ,i Y) dw
dw j du j j j
The derivative of the u function is:
du d dY
(X 2,i Y)
dw j dw j dw j

Algorithm
The derivative of the output function is:
dY

d
Y 2 2
1 dY 2 1 dY 2

1 dY 2

dw j dY 2 dw j Y 2 dw j (1 e ) dw j
Y1 2
1 dY 2 2 dY 2
2
Y

dw j dw j
J

w X
j i

1 e 0


Algorithm
The derivation of the auxiliary function Y2 gives:
w j H j
dY 2

d
dw j dY1

1 e Y1
dY1
dw j
e Y1
dY1
dw j
e
dY1
dw j
The derivative of Y1 function is:
dY1 d J

dw j dw j 0
w j H j Hj


Algorithm
Substituting the results in the equations :
J
w H
j j
j 0
e H
dE j
dw (X Y)
j dw 2, n 2
j J
w H
j j
j 0
1 e


Algorithm
The rearrangement of the output equation results in :
J
w H
j j 1 Y
j 1 1
e 1
Y Y
The final equation for the update of dwj is:
(1 Y ) H
dE j
dw (X Y)Y 2

j dw 2, n Y
j
(X Y ) Y (1 Y ) H
2, n j
Algorithm
The derivative of the hidden layer function is:
dH j

d
h2 1 dh 2 1 2 dh 2 1 h1 2 dh 2
dw j dh 2 dWi j h 2 dWi j (1 e ) dWi j
1 dh 2 2 dh 2
2
Hj

dWi j dWi j
I

Wi j X
i
1 e 0


Algorithm
The derivation of the auxiliary function h2 gives:
1 e
Xj
dh 2 d dh1 h1 dh1
ij
dh1
h1
e e
dWi j dh1 dWi j dWi j dWi j
The derivative of h1 function is:
dh1 d I W X X

dWi j dWi j 0
ij j
j

Algorithm
Substituting the results in the equations :
I
W X
ij i
e i 0 X
dE i
dW (X Y ) Y (1 Y )
ij dW 2n 2
ij I
W X
1 e i 0 i j i


Algorithm
The rearrangement of the output equation results in :
I
W X 1 H
ij i
i1 1 j
e 1
H H
j j
The final equation for the update of dWij is:
dE
dW ( X Y ) Y ( 1 Y ) H (1 H ) X
ij dW 2n j j i
ij
Algorithm
Substituting the results in the equations which is used to
iterate the wj value :
k
dE
Wikj 1 Wikj dW Wikj
ij k
dWi j
Wikj ( X k2 ,n Y k ) Y k ( 1 Y k ) H kj ( H kj 1) X ki

Derivation of Learning Algorithm
The two training algorithms are:
w k 1
j w ( X Y ) Y (1 Y ) H
k
j
k
2n
k k k k
j
Wikj 1 Wikj ( X k2 ,n Y k ) Y k ( 1 Y k ) H kj ( H kj 1) X ki
Wikj dw kj ( H kj 1) X ki

03 - Lectures of Load Forecasting - Nnat

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

03 - Lectures of Load Forecasting - Nnat

Transféré par

Droits d'auteur :

Formats disponibles

SHORT TERM LOAD FORECASTING

USING NEURAL NETWORKS AND

Arizona State University

06/2000 Short-term load forecasting 1

06/2000 Short-term load forecasting 2

06/2000 Short-term load forecasting 3

06/2000 Short-term load forecasting 4

06/2000 Short-term load forecasting 7

06/2000 Short-term load forecasting 8

This lecture presents the short term load

The short term load forecasting provides

The short term load forecasting is

06/2000 Short-term load forecasting 12

Typical short-term, daily load forecast from Salt

Typical short-term, daily load forecast from Salt

06/2000 Short-term load forecasting 14

Provide load data to the dispatchers for

The forecasted data are used for:

The forecasted data are used for:

These off-line network studies:

The system load is the sum of individual

06/2000 Short-term load forecasting 19

The factors affecting the load are:

06/2000 Short-term load forecasting 20

Economical or environmental factors

The time constraints of economical or

Time Factors affecting the load

Calls for a different model for each

Typical Seasonal Variation of Load

06/2000 Short-term load forecasting 24

Time Factors affecting the load

06/2000 Short-term load forecasting 25

Weekly Cyclic Variation

06/2000 Short-term load forecasting 26

Time Factors affecting the load

06/2000 Short-term load forecasting 27

Weather factors affecting the load

06/2000 Short-term load forecasting 28

Weather factors affecting the load

06/2000 Short-term load forecasting 29

Weather Factors Affecting the Load

06/2000 Short-term load forecasting 30

Random Disturbances Effects on Load

The different load forecasting techniques

The moving time window of data

06/2000 Short-term load forecasting 33

The selection of long periods of historical

The selection of short periods of historical

06/2000 Short-term load forecasting 34

The forecasting is a continuous process.

06/2000 Short-term load forecasting 36

The statistical methods will be discussed

This lecture concentrates on load

06/2000 Short-term load forecasting 37

06/2000 Short-term load forecasting 38

Testing the model performance

Update/modification of the model if the

06/2000 Short-term load forecasting 42

Several Artificial Neural Network (ANN)

A Feed forward Three-Layered

06/2000 Short-term load forecasting 44

The forecasting with neural network will