POWER SYSTEM DYNAMIC SECURITY ASSESSMENT – A.H.M.A.Rahim, S.K.Chakravarthy Department of Electrical Engineering K.F. University of Petroleum and Minerals.

Slides:



Advertisements
Similar presentations
1 Machine Learning: Lecture 4 Artificial Neural Networks (Based on Chapter 4 of Mitchell T.., Machine Learning, 1997)
Advertisements

Navneet Goyal, BITS-Pilani Perceptrons. Labeled data is called Linearly Separable Data (LSD) if there is a linear decision boundary separating the classes.
Support Vector Machines
Mehran University of Engineering and Technology, Jamshoro Department of Electronic Engineering Neural Networks Feedforward Networks By Dr. Mukhtiar Ali.
Software Quality Ranking: Bringing Order to Software Modules in Testing Fei Xing Michael R. Lyu Ping Guo.
Machine Learning Neural Networks
RBF Neural Networks x x1 Examples inside circles 1 and 2 are of class +, examples outside both circles are of class – What NN does.
The back-propagation training algorithm
Prénom Nom Document Analysis: Linear Discrimination Prof. Rolf Ingold, University of Fribourg Master course, spring semester 2008.
1 Hybrid Agent-Based Modeling: Architectures,Analyses and Applications (Stage One) Li, Hailin.
Development of Empirical Models From Process Data
Estimation of Oil Saturation Using Neural Network Hong Li Computer System Technology NYC College of Technology –CUNY Ali Setoodehnia, Kamal Shahrabi Department.
Hazırlayan NEURAL NETWORKS Radial Basis Function Networks I PROF. DR. YUSUF OYSAL.
Dan Simon Cleveland State University
Radial-Basis Function Networks
Face Recognition Using Neural Networks Presented By: Hadis Mohseni Leila Taghavi Atefeh Mirsafian.
Radial Basis Function Networks
Neural Networks Lecture 8: Two simple learning algorithms
1 Prediction of Software Reliability Using Neural Network and Fuzzy Logic Professor David Rine Seminar Notes.
Walter Hop Web-shop Order Prediction Using Machine Learning Master’s Thesis Computational Economics.
Radial Basis Function Networks
MSE 2400 EaLiCaRA Spring 2015 Dr. Tom Way
1/20 Obtaining Shape from Scanning Electron Microscope Using Hopfield Neural Network Yuji Iwahori 1, Haruki Kawanaka 1, Shinji Fukui 2 and Kenji Funahashi.
Artificial Neural Networks (ANN). Output Y is 1 if at least two of the three inputs are equal to 1.
Artificial Neural Networks
Using Neural Networks in Database Mining Tino Jimenez CS157B MW 9-10:15 February 19, 2009.
Artificial Neural Nets and AI Connectionism Sub symbolic reasoning.
Waqas Haider Khan Bangyal. Multi-Layer Perceptron (MLP)
11 CSE 4705 Artificial Intelligence Jinbo Bi Department of Computer Science & Engineering
1 Chapter 6: Artificial Neural Networks Part 2 of 3 (Sections 6.4 – 6.6) Asst. Prof. Dr. Sukanya Pongsuparb Dr. Srisupa Palakvangsa Na Ayudhya Dr. Benjarath.
Artificial Neural Network Supervised Learning دكترمحسن كاهاني
NEURAL NETWORKS FOR DATA MINING
 Diagram of a Neuron  The Simple Perceptron  Multilayer Neural Network  What is Hidden Layer?  Why do we Need a Hidden Layer?  How do Multilayer.
LINEAR CLASSIFICATION. Biological inspirations  Some numbers…  The human brain contains about 10 billion nerve cells ( neurons )  Each neuron is connected.
1 Adaptive Control Neural Networks 13(2000): Neural net based MRAC for a class of nonlinear plants M.S. Ahmed.
Ensembles. Ensemble Methods l Construct a set of classifiers from training data l Predict class label of previously unseen records by aggregating predictions.
1 GMDH and Neural Network Application for Modeling Vital Functions of Green Algae under Toxic Impact Oleksandra Bulgakova, Volodymyr Stepashko, Tetayna.
1 Pattern Recognition: Statistical and Neural Lonnie C. Ludeman Lecture 24 Nov 2, 2005 Nanjing University of Science & Technology.
Non-Bayes classifiers. Linear discriminants, neural networks.
Akram Bitar and Larry Manevitz Department of Computer Science
Back-Propagation Algorithm AN INTRODUCTION TO LEARNING INTERNAL REPRESENTATIONS BY ERROR PROPAGATION Presented by: Kunal Parmar UHID:
PARALLELIZATION OF ARTIFICIAL NEURAL NETWORKS Joe Bradish CS5802 Fall 2015.
1 Lecture 6 Neural Network Training. 2 Neural Network Training Network training is basic to establishing the functional relationship between the inputs.
Reservoir Uncertainty Assessment Using Machine Learning Techniques Authors: Jincong He Department of Energy Resources Engineering AbstractIntroduction.
Announcements Read Chapters 11 and 12 (sections 12.1 to 12.3)
Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.
Neural Networks Teacher: Elena Marchiori R4.47 Assistant: Kees Jong S2.22
Professor : Ming – Shyan Wang Department of Electrical Engineering Southern Taiwan University Thesis progress report Sensorless Operation of PMSM Using.
ECE 576 – Power System Dynamics and Stability Prof. Tom Overbye Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
Chapter 6 Neural Network.
1 Technological Educational Institute Of Crete Department Of Applied Informatics and Multimedia Intelligent Systems Laboratory.
1 Development of Empirical Models From Process Data In some situations it is not feasible to develop a theoretical (physically-based model) due to: 1.
IEEE AI - BASED POWER SYSTEM TRANSIENT SECURITY ASSESSMENT Dr. Hossam Talaat Dept. of Electrical Power & Machines Faculty of Engineering - Ain Shams.
Control Engineering. Introduction What we will discuss in this introduction: – What is control engineering? – What are the main types of control systems?
Lecture 2 Introduction to Neural Networks and Fuzzy Logic President UniversityErwin SitompulNNFL 2/1 Dr.-Ing. Erwin Sitompul President University
Data Mining: Concepts and Techniques1 Prediction Prediction vs. classification Classification predicts categorical class label Prediction predicts continuous-valued.
A Document-Level Sentiment Analysis Approach Using Artificial Neural Network and Sentiment Lexicons Yan Zhu.
CSE343/543 Machine Learning Mayank Vatsa Lecture slides are prepared using several teaching resources and no authorship is claimed for any slides.
EQUAL AREA STABILITY CRITERION The equal area criterion is a simple graphical method for concluding the transient stability of two-machine systems or a.
CSC 578 Neural Networks and Deep Learning
Biological and Artificial Neuron
Biological and Artificial Neuron
Biological and Artificial Neuron
Neural Network - 2 Mayank Vatsa
ECE 476 POWER SYSTEM ANALYSIS
CS621: Artificial Intelligence Lecture 22-23: Sigmoid neuron, Backpropagation (Lecture 20 and 21 taken by Anup on Graphical Models) Pushpak Bhattacharyya.
ECE 576 POWER SYSTEM DYNAMICS AND STABILITY
ECE 576 POWER SYSTEM DYNAMICS AND STABILITY
ECE 576 POWER SYSTEM DYNAMICS AND STABILITY
Akram Bitar and Larry Manevitz Department of Computer Science
Presentation transcript:

POWER SYSTEM DYNAMIC SECURITY ASSESSMENT – A.H.M.A.Rahim, S.K.Chakravarthy Department of Electrical Engineering K.F. University of Petroleum and Minerals Dhahran. CLASSICAL TO MODERN APPROACH

Background Most power systems of today have to be operated under stressed conditions. The reserve margins available and system redundancy is gradually falling Small scale disturbances may upset the power balance leading to isolation of a part of the system.

Background Large-scale disturbances can lead to breakdown of the system. Dynamic Security Assessment is related to addressing security related aspects pertaining to operation of a power system.

Dynamic Security Assessment….. Studies the combination of outages and disturbances for ensuring stable power system operation. The knowledge of operators are used as inputs to handle changing system toplogies, load pickup, sharp changes in weather etc. Based on analysis of all data, contingencies are classified to assess security and take security- related decisions.

Dynamic Security Assessment….. An important feature of a reliable power system is that it should supply power to the customers without causing disruption of service under all conditions. If a severe disturbance or fault appears on the system, it should be removed quickly enough so that the system stability under the transient condition is not affected.

Assessing Transient Stability Index The CCT is a complex function of the pre-fault and post fault system conditions, fault type, fault location, etc. and its determination is of paramount importance to power system planners. Transient stability index can be assessed by numerical integration, the second method of Lyapunov, probabilistic methods, pattern recognition, expert systems or artificial neural networks.

The Equal-Area Method The simplest way of determining the CCT is from the famous equal area criterion. Figure shows the power angle curve. The critical clearing time t cr corresponds to angle  cr when area A 1 equals A 2.

Numerical Integration The most widely used method of determining CCT. The system dynamics is solved repetitively until the fault duration is obtained which takes the system to the threshold of instability. For accurate results the step size should small, making the process very slow.

Lyapunov Method In the second method of Lyapunov the post- fault system equations are replaced by a stability criterion. The figure shows the stable and unstable manifolds in the  -  phase plane.

Probabilistic Method Generally, the determination of CCT through these methods is quite involved. The probabilistic methods attempt to assess the probability of stability of a system following exposure to a disturbance. In some studies, probability functions, called the security function, are compared with a maximum tolerable insecurity or risk level to determine if and when some control action has to be taken.

Some Modern Methods Pattern recognition based transient stability studies, attention have been focused on selection of the initial system description, feature extraction and classifier design. The expert system approach combines the time domain numerical approach with human expert knowledge coded in a rule based program. The methods used in artificial neural network employ adaptive pattern recognition approach, which trains the neurons to learn from the sets of input-output presented to it.

Application of ANN in determining CCT For the network shown faults were simulated on buses 2 and 3. Fault-clearing policy is to restore pre-fault topology. Simulations were also made with line 2-5 being removed. Twenty different loading conditions were considered G G G G

Back-propagation type ANN In the training process, the network is presented with a pair of patterns – an input pattern and a corresponding desired output pattern. In the forward pass, the outputs are computed on the basis of selected weights and the error is computed. In the backward pass the weights are updated so as to minimize the sum of the squares of errors, bpbp y1y1 ymym [w rm ] b1b1 x1x1 xpxp v1v1 vrvr [w pr ]

Radial Basis Type ANN In training the RBF network the linear weights are estimated so as to minimize the sum of the square of the error between the desired output d(k) and the network output y(k). An orthogonal least squares approach chooses the centers of the radial basis functions as subsets of the weighting matrix from a linear regression model of the error equation. This method has been employed in this study. y1y1 ymym w kj x1x1 xpxp c1c1

Training data The rotor angle of each machine relative to the center of inertia    at the instant of fault initiation Accelerating power of each generator immediately following the fault Accelerating kinetic energy of each machine given as x i =      i=1,2,…4 x i+4 = P mi - P ei, i=1,2,…4 x i+8 = (P mi -P ei )/ H i, i=1,2,….4

Results: Convergence Characteristics

Results: Comparison

Comparison of CCT

Conclusions A relatively modern method of estimating CCT through artificial neural networks has been presented. Both back-propagation as well as the radial-basis function ANN’s were trained to estimate the critical fault clearing times of a multi-machine power system. It was observed that the RBF networks estimates the CCT with better accurately.

Conclusions THANK YOU FOR YOUR ATTENTION