Electric lectrical al load forecasting load forecasting E using - PowerPoint PPT Presentation

Electric lectrical al load forecasting load forecasting E using artificial neural using artificial neural network kohonen kohonen network methode methode Galang Jiwo Syeto / EEPIS- Galang Jiwo Syeto / EEPIS -ITS ITS 7406.040.058 7406.040.058

INTRODUCTION INTRODUCTION � Electricity can not be stored in a large scale, Electricity can not be stored in a large scale, � therefore this power must be provided when therefore this power must be provided when needed. needed. � As a result there is a problem in As a result there is a problem in unfixed unfixed � electrical power quota quota, how to operate an , how to operate an electrical power electric power system that that always able to meet always able to meet electric power system the power demand at any time, in a in a good quality. good quality. the power demand at any time,

INTRODUCTION INTRODUCTION � The first prerequisite should be implemented to The first prerequisite should be implemented to � achieve the the goal goal, ,the electric company the electric company must must achieve knows the electrical electrical load or power demand in load or power demand in knows the the future. the future. � So that So that � � s why we need, ELECTRICAL LOAD s why we need, ELECTRICAL LOAD � FORECASTING FORECASTING

Final project Objectives Final project Objectives � Build Build electrical load forecasting system electrical load forecasting system more more � accurate in in minimum minimum average average erro error r accurate � Compare hybrid backpropagation kohonen and Compare hybrid backpropagation kohonen and � hybrid counterpropagation kohonen in counterpropagation kohonen in hybrid electrical forecasting system electrical forecasting system

Problems Problems How to determine algorithm using How to determine algorithm using � � backpropagation with kohonen and backpropagation with kohonen and counterpropagation with kohonen for counterpropagation with kohonen for electrical forecasting and get minimum error electrical forecasting and get minimum error How to determine the number of hidden layer hidden layer How to determine the number of � � in backpropagation and counter propagation in backpropagation and counter propagation methode methode

Limitations Issue Limitations Issue Used sed Artificial Neural Network Artificial Neural Network especially especially U � � backpropagation,counterpropagation and backpropagation,counterpropagation and kohonen . . kohonen Input data used is the electrical load data taken Input data used is the electrical load data taken � � from PLN PLN Company, Company, Channeling Channeling and Load and Load from Management Center Center division for East Java division for East Java and and Management st Bali between September,1 ,1 st 2005 until Bali between September 2005 until th 200 January,30 th 2006 6 January,30 Static input data in .txt file Static input data in .txt file � �

System design (ANN Architecture) System design (ANN Architecture) � Hybrid methode backpropagation Hybrid methode backpropagation- -kohonen kohonen � � 2 node in input layer,4 node in hidden layer,2 2 node in input layer,4 node in hidden layer,2 � node in output layer (BP) node in output layer (BP) � 122 node in input layer and 122 node in output 122 node in input layer and 122 node in output � layer kohonen layer kohonen

System Design (General Method) System Design (General Method) � First,we calculate mean and standard deviation First,we calculate mean and standard deviation � each day each day � Second,we calculate normalization profile each Second,we calculate normalization profile each � day day � Third, Get the prediction for the mean and Third, Get the prediction for the mean and � standard deviation for next day standard deviation for next day � Get the prediction Get the prediction �

System design (ANN Architecture) System design (ANN Architecture) � Hybrid methode counterpropagation Hybrid methode counterpropagation- -kohonen kohonen � � 2 node in input layer,4 node in hidden layer,2 2 node in input layer,4 node in hidden layer,2 � node in output layer (CP) node in output layer (CP) � 122 node in input layer and 122 node in output 122 node in input layer and 122 node in output � layer kohonen layer kohonen

System Design (Data) System Design (Data) � Data for this system Data for this system was was electrical load data electrical load data � taken from PLN PLN Company, Company, Channeling Channeling and and taken from Load Management Center Center division for East Java division for East Java Load Management and Bali between between September,1 ,1 st st 2005 until until and Bali September 2005 th January,30 anuary,30 th 2006 per hour in 6 per hour in mega- -watt watt J 200 mega units,total of the data are 3648 data ,total of the data are 3648 data units

System Design (Data) System Design (Data) PERIODE INPUT PERIODE INPUT PERIODE INPUT PERIODE INPUT PERIODE INPUT PERIODE INPUT PERIODE INPUT PERIODE OUTPUT PERIODE OUTPUT PERIODE OUTPUT PERIODE OUTPUT PERIODE OUTPUT PERIODE OUTPUT PERIODE OUTPUT DATA TRAINING DATA TRAINING DATA TRAINING DATA TRAINING DATA TRAINING DATA TRAINING DATA TRAINING DATA TES DATA TES DATA TES DATA TES DATA TES DATA TES DATA TES

Implementation (BP initialization) Implementation (BP initialization) Initial 3 weight randomize between 0 0 until until 1 1 Initial 3 weight randomize between � � Initial Alpha for backpropagation Initial Alpha for backpropagation � � Maximum error value Maximum error value � � Sigmoid function value (lambda) Sigmoid function value (lambda) � � Initial Alpha for kohonen Initial Alpha for kohonen � � Epoch value for kohonen Epoch value for kohonen � �

Implementation (CP initialization) Implementation (CP initialization) Initial 3 weight randomize between 0 0 until until 1 1 Initial 3 weight randomize between � � Initial Learning Rate alpha and beta Initial Learning Rate alpha and beta � � width neighbors neighbors controller controller Fun Funct cti ion on (k0), (k0), width � � Number of counterpropagation Epoch Number of counterpropagation Epoch � � Initial Alpha for kohonen Initial Alpha for kohonen � � Epoch value for kohonen Epoch value for kohonen � �

Implementation (data preprocessing) Implementation (data preprocessing) � Normalization Normalization � � Mean and deviation standart Mean and deviation standart �

Implementation (data preprocessing) Implementation (data preprocessing) � Normalization Profil Normalization Profil �

Implementation (NN Training) Implementation (NN Training) � Backpropagation Backpropagation � � Counterpropagation Counterpropagation � � Get the best weight for the mean and standar Get the best weight for the mean and standar � deviation forecasting deviation forecasting

Implementation (Classification) Implementation (Classification) � Classify mean and deviation standart in 122 Classify mean and deviation standart in 122 � group classification group classification � Classify mean and deviation standart result of Classify mean and deviation standart result of � the forecasting the forecasting � Get the normalization profil index Get the normalization profil index �

Implementation (Postprocessing) Implementation (Postprocessing) � Use euclidiance distance to get the nearest Use euclidiance distance to get the nearest � normalization profil index normalization profil index

Implementation (Forecasting Result) Implementation (Forecasting Result) � Get the forecasting result using Get the forecasting result using �

Implementation (Mean Square Error) (Mean Square Error) Implementation � To assess the performance of this forecasting To assess the performance of this forecasting � system, we use MSE for getting we use MSE for getting error error system, calculation calculation � MSE = � 2 /n (Actual i - Fitted Fitted i ) 2 /n MSE = (Actual i - i )

Testing and Analyzation Testing and Analyzation � Determine the number of neuron in hidden layer Determine the number of neuron in hidden layer � (backpropagation- -kohonen) kohonen) (backpropagation Jumlah Neuron MSE Training MSE Peramalan 2 0,000007 239242,8067 2 0,000007 239242,8067 3 0,000012 235178,5398 3 0,000012 235178,5398 4 0,000021 230566,4836 4 0,000021 230566,4836 5 0,000006 240897,7882 5 0,000006 240897,7882 6 6 0,000002 0,000002 248966,8643 248966,8643

Testing and Analyzation Testing and Analyzation � Determine the number of neuron in hidden layer Determine the number of neuron in hidden layer � (counterpropagation- -kohonen) kohonen) (counterpropagation Jumlah Neuron MSE Training MSE Peramalan 2 0,013299 382313,4391 3 0,069946 451102,7709 4 0,094297 2010904,1890 5 0,078252 1714283,0143 6 0,079253 10418776,1022

Testing and Analyzation Testing and Analyzation � Compare forecasting value hybrid from Compare forecasting value hybrid from � backpropagation kohonen and counterpropagation backpropagation kohonen and counterpropagation kohonen kohonen � One next step forecasting One next step forecasting � metode MSE CPNN-Kohonen 406242,4146 BPNN-Kohonen 893066,4288