Tsoukalas Purdue UTSA DDDAS JAN 2016

EAGER – Dynamic Data: Machine Intelligence for Dynamic DataDriven Morphing of Nodal Demand in Smart Energy Systems CO-PRINCIPAL INVESTIGATORS: PRINCIPAL INVESTIGATOR:

Prof. Lefteri H. Tsoukalas Purdue University

- Prof. Nikolaos Gatsis University of Texas at San Antonio - Prof. Miltos Alamaniotis Purdue University

2016 PI meeting of the AFOSR DDDAS Program and the joint NSF/AFOSR DDS Program

Outline  

Smart Energy Systems Nodal Demand Morphing  Big

 

Data Problem

Proposed Approach Virtual Budget  Results



Summary / Questions

Lefteri H. Tsoukalas, AFOSR 2016

2

Smart Energy Systems ANTICIPARION

Main Components

1

2

3

DATA ANALYTICS Machine Intelligence / Signal Processing

COMPUTING & COMMUNICATIONS

DATA

ENERGY SYSTEMS

Lefteri H. Tsoukalas, AFOSR 2016

3

Problem: Nodal Electricity Morphing as Big Data Problem  



Model the electric grid as demand-driven system Information and price signals dynamically vary with time and demand

Big Data: Diverse and complex infrastructure of energy systems  Demanding



and data intensive activities

Big Data Challenge: Utilization of real-time load decision-making based on dynamic information Lefteri H. Tsoukalas, AFOSR 2016

4

Proposed Approach 

Dynamic Data Driven System Application (DDDAS) in Load Morphing



Interaction of Machine Learning with Dynamic Optimization Lefteri H. Tsoukalas, AFOSR 2016

5

Virtual Budget: Automated Load Morphing Method Elastic and Inelastic Load

Machine Learning

Dynamic Optimization

Lefteri H. Tsoukalas, AFOSR 2016

6

Virtual Budget: Optimization Problem and Decision Making Optimization Problem

Lefteri H. Tsoukalas, AFOSR 2016

7

Preliminary Results Virtual Budget Levels

User Input: Max Cost Threshold

Lefteri H. Tsoukalas, AFOSR 2016

8

Computed Costs

Lefteri H. Tsoukalas, AFOSR 2016

9

Summary & Future Work 

Smart Energy Systems  Dynamic

information from multiple sources  Big data problem 

Integration of:  Machine

learning  Dynamic optimization 

Future Work  Simulation

platforms  Acceleration of algorithms  Extensive testing Lefteri H. Tsoukalas, AFOSR 2016

10

Publications Journal Paper [1] Alamaniotis, M., Bargiotas, D., & Tsoukalas, L.H., “Towards Smart Energy Systems: Application of Kernel Machine Regression for Medium Term Electricity Load Forecasting,” SpringerPlus – Engineering Section, Springer, Vol 5, (1), pp. 1-15, 2016.

Conference Proceeding Paper [1] Gatsis, N., Yalamanchili, L., Bazrafshan, M., & Risbud, P., “Decentralized Coordination of Energy Resources in Electricity Distribution Networks,” in Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing, Shangai, China, Apr. 2016. [2] Alamaniotis, M., Bourbakis, N., & Tsoukalas, L.H., “Very-Short Term Forecasting of Electricity Price Signals Using a Pareto Composition of Kernel Machines in Smart Power Systems,” 3rd IEEE Global Conference on Signal and Information Processing,, pp. 1-5, Orlando, FL, December 2015.

Journal Paper Submission [1] Alamaniotis, M., Gatsis, N., & Tsoukalas, L.H., “Virtual Budget: Integration of Electricity Load and Price Anticipation for Load Morphing in Price-Directed Energy Utilization,” IEEE Transactions on Power Systems, Institute of Electrical and 11 Electronic Engineers. Lefteri H. Tsoukalas, AFOSR 2016

Thanks for your Attention Questions? Comments?

Lefteri H. Tsoukalas, AFOSR 2016

12