REAL-TIME PRICE FORECAST WITH BIG DATA



Similar documents
Short Term Electricity Price Forecasting Using ANN and Fuzzy Logic under Deregulated Environment

Artificial Neural Network-based Electricity Price Forecasting for Smart Grid Deployment

Probabilistic Forecasting of Medium-Term Electricity Demand: A Comparison of Time Series Models

Real-time Security-Constrained Economic Dispatch and Commitment in the PJM : Experiences and Challenges

Sales Forecast for Pickup Truck Parts:

Grid Operations and Planning

Energy in the Wholesale Market December 5, :30 p.m. to 3:30 p.m. Irving, Texas Robert Burke, Principal Analyst ISO New England

Hybrid processing of SCADA and synchronized phasor measurements for tracking network state

Convergence Bidding Tutorial & Panel Discussion

Overview and Comparison of Demand Response Programs in North American Electricity Markets

ECCO International, Inc. 268 Bush Street, Suite 3633 San Francisco, CA 94104

Modeling and Prediction of Network Traffic Based on Hybrid Covariance Function Gaussian Regressive

2Azerbaijan Shahid Madani University. This paper is extracted from the M.Sc. Thesis

Curriculum Vitae July 2011 Shu Fan

Analysis of Financial Time Series

Course notes for EE394V Restructured Electricity Markets: Locational Marginal Pricing

Java Modules for Time Series Analysis

Gaussian Processes to Speed up Hamiltonian Monte Carlo

ADVANCED LOCAL PREDICTORS FOR SHORT TERM ELECTRIC LOAD

INTERNATIONAL JOURNAL OF PURE AND APPLIED RESEARCH IN ENGINEERING AND TECHNOLOGY

CLOUD DATABASE ROUTE SCHEDULING USING COMBANATION OF PARTICLE SWARM OPTIMIZATION AND GENETIC ALGORITHM

International Journal of Software and Web Sciences (IJSWS)

APPLICATION OF ADVANCED SEARCH- METHODS FOR AUTOMOTIVE DATA-BUS SYSTEM SIGNAL INTEGRITY OPTIMIZATION

Modeling and Forecasting Day-Ahead Hourly Electricity Prices: A Review

Information Theory and Stock Market

Large-Scale IP Traceback in High-Speed Internet

Enabling 24/7 Automated Demand Response and the Smart Grid using Dynamic Forward Price Offers

Optimization Models for Advanced Cyber-security in the Smart Grid. Seyedamirabbas Mousavian

Microgrid: A new hub in energy infrastructure. Mohammad Shahidehpour Illinois Institute of Technology

Power System Security Monitoring, Analysis, and Control. George Gross

AN EFFICIENT STRATEGY OF AGGREGATE SECURE DATA TRANSMISSION

A Time Series Analysis of the Chicago Board Options Exchange Volatility Index (VIX)

Bayesian Machine Learning (ML): Modeling And Inference in Big Data. Zhuhua Cai Google, Rice University

11. Time series and dynamic linear models

Stochastic Security-Constrained Unit Commitment Lei Wu, Mohammad Shahidehpour, Fellow, IEEE, and Tao Li, Member, IEEE

Droop Control Forhybrid Micro grids With Wind Energy Source

Mata : Garuda An advanced Network Monitoring System The S.L.A.D Network Security Framework. FIRST Conference Berlin, 19 June 2015

Time Series Analysis of Household Electric Consumption with ARIMA and ARMA Models

Detection. Perspective. Network Anomaly. Bhattacharyya. Jugal. A Machine Learning »C) Dhruba Kumar. Kumar KaKta. CRC Press J Taylor & Francis Croup

Capacity Limits of MIMO Channels

A Comparative Study of the Pickup Method and its Variations Using a Simulated Hotel Reservation Data

Fuzzy Cognitive Map for Software Testing Using Artificial Intelligence Techniques

Learning outcomes. Knowledge and understanding. Competence and skills

Assessment Plan for CS and CIS Degree Programs Computer Science Dept. Texas A&M University - Commerce

Manjeet Kaur Bhullar, Kiranbir Kaur Department of CSE, GNDU, Amritsar, Punjab, India

Introduction to Markov Chain Monte Carlo

Wind Power and Electricity Markets

PJM Overview of Markets. Georgian Delegation PUCO Office April 11, 2013

ERCOT Monthly Operational Overview (March 2014) ERCOT Public April 15, 2014

Generating Valid 4 4 Correlation Matrices

A New Quantitative Behavioral Model for Financial Prediction

Performance Analysis of Data Mining Techniques for Improving the Accuracy of Wind Power Forecast Combination

Section 4: Scheduling Philosophy & Tools

An Introduction to Variable-Energy-Resource Integration Analysis Energy Exemplar October 2011

ENERGY DEMAND MODEL DESIGN FOR FORECASTING ELECTRICITY CONSUMPTION AND SIMULATING DEMAND RESPONSE SCENARIOS IN SWEDEN

A Wavelet Based Prediction Method for Time Series

Uncertainty modeling revisited: What if you don t know the probability distribution?

Assessment of Price Risk of Power under Indian Electricity Market

Real-Time Monitoring of Complex Industrial Processes with Particle Filters

Tutorial on Markov Chain Monte Carlo

Demand Forecasting Optimization in Supply Chain

How to Detect and Prevent Cyber Attacks

itesla Project Innovative Tools for Electrical System Security within Large Areas

Discrete Frobenius-Perron Tracking

International Journal of Emerging Technology & Research

Monte Carlo testing with Big Data

SMART ENERGY SMART GRID. More than 140 Utilities companies worldwide make use of Indra Solutions. indracompany.com

A New Method for Electric Consumption Forecasting in a Semiconductor Plant

Supply Chain Forecasting Model Using Computational Intelligence Techniques

COPYRIGHTED MATERIAL. Contents. List of Figures. Acknowledgments

Energy Demand Forecasting Industry Practices and Challenges

Unlocking Electricity Prices:

Stock Market Forecasting Using Machine Learning Algorithms

Machine Learning Techniques for Stock Prediction. Vatsal H. Shah

Application of linear programming methods to determine the best location of concrete dispatch plants

Risk Management for IT Security: When Theory Meets Practice

Power System review W I L L I A M V. T O R R E A P R I L 1 0,

Dynamic Thermal Ratings in Real-time Dispatch Market Systems

Non Linear Dependence Structures: a Copula Opinion Approach in Portfolio Optimization

Pontifical Catholic University of Parana Mechanical Engineering Graduate Program

ENHANCED HYBRID FRAMEWORK OF RELIABILITY ANALYSIS FOR SAFETY CRITICAL NETWORK INFRASTRUCTURE

Video Game Description Language (VGDL) and the Challenge of Creating Agents for General Video Game Playing (GVGP)

Introduction to the Integrated Marketplace

AN EFFICIENT LOAD BALANCING ALGORITHM FOR A DISTRIBUTED COMPUTER SYSTEM. Dr. T.Ravichandran, B.E (ECE), M.E(CSE), Ph.D., MISTE.,

THE LEAN-RESOURCES BASED CONSTRUCTION PROJECT PLANNING AND CONTROL SYSTEM

Neural Network Based Forecasting of Foreign Currency Exchange Rates

Integrated Resource Plan

A Novel Binary Particle Swarm Optimization

Electrical Power and Energy Systems

Transcription:

REAL-TIME PRICE FORECAST WITH BIG DATA A STATE SPACE APPROACH Lang Tong (PI), Robert J. Thomas, Yuting Ji, and Jinsub Kim School of Electrical and Computer Engineering, Cornell University Jie Mei, Georgia Institute of Technology August 7, 2013 CERTS Review

DATA QUALITY AND ITS EFFECTS ON MARKET OPERATIONS DENY OF SERVICE ATTACK ON REAL-TIME ELECTRICITY MARKET COVER UP PROTECTION AGAINST TOPOLOGY ATTACK Lang Tong (PI), Robert J. Thomas, and Jinsub Kim Cornell University August 8, 2013 CERTS Review

Project overview Objectives Accurate short-term probabilistic forecasting of real-time LMP. Incorporate real-time measurements (e.g. SCADA/PMU). Scalable computation techniques. Summary of results A real-time LMP model with forecasting and measurement uncertainties. A Markov chain abstraction of real-time LMP computation. Monte Carlo sampling techniques. Preliminary simulations.

Outline Motivation Load and real-time LMP as random processes Benchmark techniques Comments on the state of the art. Probabilistic forecasting of real-time LMP Simulation studies Summary and future work

Sample paths of load and LMP

Day ahead vs. real-time

Benchmark techniques Time series ARMA, ARIMA, ARMAX, GARCH Machine learning Neural networks, support vector machines (SVM) Hybrid techniques Jump (switching) models

Load forecasting Forecasting Method ANN [1] ANN [2] MAPE 1.46% 1.24% SVM [3] 2.09% [1] P. Mandal, T. Senjyu, N. Urasaki, and T. Funabashi, A neural network based several-hour-ahead electric load forecasting using similar days approach, Int. Journal. of Electric Power and Energy System, vol. 28, no. 6, pp. 367 373, July 2006. [2] P. Mandal, T. Senjyu, N. Urasaki, and T. Funabashi, Short-Term Price Forecasting for Competitive Electricity Market, Power Symposium, 2006. NAPS 2006. 38 th North. [3] S. Fan, C. Mao and L. Chen, Next-day electricity-price forecasting using a hybrid network, IET Generation, Transmission & Distribution, Volume 1, Issue 1, January, 2007.

Day ahead LMP forecasting Forecasting Method ANN [1] ANN [2] ANN [4] MAPE 15.96% 12.92% 8.67% SVM [3] 11.31% GARCH [5] 11.55% ARMAX [6] 9.01% [4] P. Mandal, T. Senjyu, N. Urasaki, and T. Funabashi, A neural network based several-hour-ahead electric load forecasting using similar days approach, Int. Journal. of Electric Power and Energy System, vol. 28, no. 6, pp. 367 373, July 2006. [5] R. Garcia, J. Contreras, A GARCH Forecasting Model to Predict Day-Ahead Electricity Prices, Int. Journal of Electric Power and Energy System, vol. 20, no. 2, MAY 2005. [6] J. Zhang, C. Yu, G. Hou, Application of Chaotic Particle Swarm Optimization in the Short-term Electricity Price Forecasting, Int. Journal of Electric Power and Energy System, vol. 28, no. 6, pp.

Real-time LMP forecasting Actual and forecast six-hours-ahead electricity price for Victorian electricity market (Nov 1 st to 7 th, 2003). Actual and forecast six-hours-ahead electricity price for PJM market (Jan 8 th to 14 th, 2006). [4] F. Sarafraz, H. Ghasemi, H. Monsef, Locational Marginal Price Forecasting by Locally

Summary of the state-of-the-art Extensive literature: Mostly black box techniques Primarily providing point forecasting Rarely deal with on LMP and network effects Extremely accurate load forecasting (1-3%) Relatively poor price forecasting (10-20%)

Outline Motivation Probabilistic forecasting of real-time LMP A stylized real-time ex post LMP model Information structure LMP states and Markov chain representations Probabilistic forecast Preliminary simulations Summary and future work

A stylized real-time ex post LMP model

State estimation

Ex ante dispatch

Ex post eligible generators

Ex post LMP

Ex post LMP via IDCOPF

Information structure and Markov chain

Probabilistic forecasting

Outline Motivation Probabilistic forecasting of real-time LMP Preliminary simulations IEEE 16 bus with NYISO sample load profiles Varying load errors and correlations Monte Carlo techniques to estimate transition matrices Summary and future work

Load and price distributions 1 LMP distribution at time 15 38 36 34 Probability 0.5 0 1 2 3 4 5 6 7 8 9 1011 Markov State (LMP) index Scaled average actual load profile (upper/lower bounds for predicted load profile) from NYISO June 1 st actual load profile upper bound lower bound 32 1 LMP distribution at time 42 MW 30 28 26 24 22 Probability 0.5 0 1 2 3 4 5 6 7 8 9 1011 Markov State (LMP) index Probability 1 0.5 0 LMP distribution at time 119 1 2 3 4 5 6 7 8 9 1011 Markov State (LMP) index Probability 1 0.5 0 LMP distribution at time 220 1 2 3 4 5 6 7 8 9 1011 Markov State (LMP) index 20 0 15 42 119 200 288 time index (5 mins interval in 24 hours)

LMP trajectories $/MW 50 45 40 35 30 25 load bus1 bus2 bus3 bus4 bus5 bus6 bus7 bus8 bus9 bus10 bus11 bus12 bus13 bus14 LMP trajectories 20 15 0 50 100 150 200 250 time index (5 mins interval in 24 hours)

Time inhomogeneous Markov Chain

Transition Matrices Starting Markov State (LMP) index Transition matrix at time 15 2 4-1 6-2 8 10-3 12-4 2 4 6 8 10 12 Ending Markov State (LMP) index Starting Markov State (LMP) index 2 4 6 8 10 12 Transition matrix at time 29 2 4 6 8 10 12 Ending Markov State (LMP) index -1-2 -3 Starting Markov State (LMP) inde Transition matrix at time 117 2 0 4-2 6-4 8 10-6 12-8 2 4 6 8 10 12 Ending Markov State (LMP) index Starting Markov State (LMP) index 2 4 6 8 10 12 Transition matrix at time 220 2 4 6 8 10 12 Ending Markov State (LMP) index 0-1 -2-3

6 hours ahead prediction error Probability of incorrect prediction Absolute Percentage Error (MAPE 10 0 10-2 10-4 10-6 10 8 6 4 2 0 Prediction error for best and second best prediction Best prediction 2 nd best prediction 20 40 60 80 100 120 140 160 180 200 time index (5 mins interval in 12 hours) Starting Markov State (LMP) index Prediction transition matrix at time 21 0 2-0.5 4-1 6-1.5 8-2 10-2.5 12-3 2 4 6 8 10 12 Ending Markov State (LMP) index MAPE of best (most likely) prediction 20 40 60 80 100 120 140 160 180 200 time index (5 mins interval in 12 hours) Starting Markov State (LMP) index Prediction transition matrix at time 21 0 2-0.5 4-1 6-1.5 8-2 10-2.5 12-3 2 4 6 8 10 12 Ending Markov State (LMP) index

LMP prediction for different time horizon $/MW 50 40 30 20 Actual Best prediction 2 nd best LMP prediction for 5 mins ahead $/MW 50 40 30 20 LMP prediction for 1 hour ahead Actual Best prediction 2 nd best 10 10 $/MW 0 50 40 30 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Bus index Actual Best prediction 2 nd best LMP prediction for 3 hours ahead $/MW 0 50 40 30 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Bus index LMP prediction for 6 hours ahead Actual Best prediction 2 nd best 10 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Bus index 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Bus index

LMP distribution for different prediction horizons LMP distribution for 5 mins ahead LMP distribution for 1 hour ahead 10 0 10 0 Probability (Log) 10-2 Probability (Log) 10-2 10-4 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Markov State (LMP) index LMP distribution for 3 hours ahead 10-4 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Markov State (LMP) index LMP distribution for 6 hours ahead Probability (Log) 10-2 Probability (Log) 10-2 10-4 1 2 3 4 5 6 7 8 9 10 11 12 13 Markov State (LMP) index 10-4 1 2 3 4 5 6 7 8 9 10 11 12 13 Markov State (LMP) index

Planned research Forecasting techniques Incorporating load models More sophisticated Monte Carlo techniques (important sampling, MCMC) Generator behavior models Forecasting with renewable sources Demand response Extensive simulation studies and comparisons

DATA QUALITY AND ITS EFFECTS ON MARKET OPERATIONS DENY OF SERVICE ATTACK ON REAL-TIME ELECTRICITY MARKET Lang Tong (PI), Robert J. Thomas, and Jinsub Kim Cornell University Presented at CERTS Review August 7-8, 2013

Observability

A graph theoretic condition (KCD 80)

Undetectable attack (KJTT 11)

Only half of the meters are needed!

Bad data identification and removal

Framing attack via QCQP

DoS attack on real-time operations

Simulation examples (AC)

Conclusion Attacks on cyber physical systems (power systems) are real: State attacks Topology attacks Dispatch attacks A key step toward protection is deploy authentication mechanisms.