Authors

Muqi Li

Type

Text

Type

Dissertation

Advisor

Feinberg, Eugene | Samulyak, Roman | Hu, Jiaqiao | Robertazzi, Thomas.

Date

2015-12-01

Keywords

Electrical engineering | Load Forecasting, PMU Placement, Power Flow Calculation, Smart Grid

Department

Department of Applied Mathematics and Statistics.

Language

en_US

Source

This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree.

Identifier

http://hdl.handle.net/11401/76018

Publisher

The Graduate School, Stony Brook University: Stony Brook, NY.

Format

application/pdf

Abstract

This dissertation studies two problems in smart grid, one of which is real-time power flow analysis. Power flow analysis is used to obtain the steady-state voltage phasors for the power system. The ability to perform power flow analysis quickly is essential for the successful implementation of advanced real-time control of transmission systems. We describe a sensor placement algorithm for conducting real-time parallel transmission network power flow computations. In particular, Phasor Measurement Units (PMUs) can be such sensors. Graph partitioning is used to decompose the system into several subsystems and to locate sensors in an efficient way. Power flow calculations are then run in parallel for each area. Test results on the IEEE 118- and 300-bus systems show that the proposed algorithm is faster than the traditional(serial) Newton’s method, and is suitable for real-time applications. Electricity load forecasting is another problem investigated in this dissertation. Electric load forecasting techniques are used by most electric utility companies for operation and planning. Many operational and financial decisions are based on load forecasting, such as reliability analysis, voltage control,unit commitment,security assessment, and in purchasing electric power. We focus on short-term electric load forecasting. For this problem we present two models that predict future electricity demands based on historical hourly load and hourly weather information. A data cleaning scheme is applied to make the models robust. The estimation of the next day load is performed with an Artificial Neural Network (ANN) method and a Modified Statistical Learning method (MSL). We compare the results obtained by ANN and MSL method. Numerical testing shows that both methods provide accurate predictions. | This dissertation studies two problems in smart grid, one of which is real-time power flow analysis. Power flow analysis is used to obtain the steady-state voltage phasors for the power system. The ability to perform power flow analysis quickly is essential for the successful implementation of advanced real-time control of transmission systems. We describe a sensor placement algorithm for conducting real-time parallel transmission network power flow computations. In particular, Phasor Measurement Units (PMUs) can be such sensors. Graph partitioning is used to decompose the system into several subsystems and to locate sensors in an efficient way. Power flow calculations are then run in parallel for each area. Test results on the IEEE 118- and 300-bus systems show that the proposed algorithm is faster than the traditional(serial) Newton’s method, and is suitable for real-time applications. Electricity load forecasting is another problem investigated in this dissertation. Electric load forecasting techniques are used by most electric utility companies for operation and planning. Many operational and financial decisions are based on load forecasting, such as reliability analysis, voltage control,unit commitment,security assessment, and in purchasing electric power. We focus on short-term electric load forecasting. For this problem we present two models that predict future electricity demands based on historical hourly load and hourly weather information. A data cleaning scheme is applied to make the models robust. The estimation of the next day load is performed with an Artificial Neural Network (ANN) method and a Modified Statistical Learning method (MSL). We compare the results obtained by ANN and MSL method. Numerical testing shows that both methods provide accurate predictions. | 102 pages

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.