Abstract: In a deregulated electricity market, the financial transmission rights (FTRs) and the bid-sell principle for energy trades are used to determine the expected power flows on transmission lines. Expected power flows are calculated by applying the superposition theorem on the approved electronic tags (e-tags). Multiple parallel paths in interconnected networks lead to division of power flows determined by the impedances of the parallel paths and the physical laws of electricity. The actual power flows in the network do not conform to the market expectations leading to unscheduled flows (USF) on transmission lines. USF have historically been estimated and accommodated in a deterministic sense for a given set of e-tags. However, wide-area interconnections experience variability and uncertainty due to a significant penetration of wind energy connected at the transmission level, thus imparting a stochastic nature to USF. This research develops an automated framework that provides a market level accommodation of variable USF. This generic framework will be applicable to any transmission network with intermittent energy resources. A linear model, from the literature, has been adopted to model USF using a mathematical artifact called minor loop flows. Techniques to automate the elements of the linear estimator are developed. A loop detection algorithm (LDA) based on graph theory is proposed to detect loops in a transmission network of any size. The LDA is formulated as a modification of the A-star (A*) algorithm, the lowest ancestor theorem, and Dijkstras algorithm. The LDA has an order of complexity of (V)2, where V is the total number of vertices or buses in the network under consideration. An application of a geographical information systems (GIS) technique has been established to obtain the transmission line layouts. The outcome of the LDA (i.e., minor loops) and line layouts (i.e., azimuth) are processed to compute the incidence matrix of the estimator. The variability due to the penetration of wind energy is accounted in the proposed framework using the probabilistic load flow analysis based on Monte Carlo simulations. Suitable estimation techniques are applied to solve the linear estimator to obtain estimates of variable minor loop flows. Two techniques - ordinary least squares (OLS) and analytic ridge regression (RR) - are used to estimate minor loop flows. The estimation techniques adhere to the auto-correction of the quality of estimates in case of ill-conditioning of the incidence matrix. Wind power generation companies (WGENCOs) employ forecasting models to participate in the primary electricity markets. Forecasting models used to predict the output of wind plants are inherently erroneous and hence their impacts on USF are studied. The impact of forecasting errors associated with the output of wind plants is investigated using the concept of prediction intervals rather than point forecast values. The proposed framework is tested on the standard IEEE 14 bus and 30 bus test networks. Accurate loops are detected for the aforementioned networks using the LDA. Loops detected by the LDA and GIS coordinates of the IEEE 14 bus test system are used to demonstrate the synthesis of incidence matrix. An annual intermittency case of generation resource is simulated using the measurements of wind plant penetrations to accommodate variable USF. Loop flow estimates are computed using OLS and analytic RR (to counter multi-collinearity of incidence matrix). Prediction intervals for the output of wind power plants are computed using the forecasting error distribution normal distribution. Future work involves formulation of the mapping of e-tag components (point of receipt/point of delivery) to determine market expectations used in electricity trades. Furthermore, a contribution factor for wind power plants towards variable USF in the network will also be developed. Thus, an automated and generic market level accommodation of variable USF due to generation resource intermittency will be accomplished.

Adviser: Dr. Siddharth Suryanarayanan

Co-Adviser: N/A

Non-ECE Member: Prof. Daniel Zimmerle

Member 3: Dr. Peter Young

Addional Members: Dr. Liuqing Yang

Publications:

1. M. Mohanpurkar, S. Suryanarayanan, "Accommodating Unscheduled Flows in Electric Grids Using the Analytical Ridge Regression", in press, IEEE Trans. on Power Systems, 2 pp.

2. M. Mohanpurkar, H. Valdiviezo, and S. Suryanarayanan, "An Application of Geographical Information Systems Technique in the Estimation of Loop Flows in Wide Area Transmission Networks with High Wind Penetration," 45th Proc. of Frontiers of Power Conference, Oklahoma, October 2012, pp. VI-1 - VI-8.

3. M. Mohanpurkar, S. Suryanarayanan, "A Case Study on the Effects of Predicted Wind Farm Power Outputs on Unscheduled Flows in Transmission Networks," 5th Annual Green Technologies Conference, April 2013, Colorado. (Accepted).

4. M. Connell, M. Mohanpurkar, D. Zimmerle, "Characterization of Power and Energy Storage Requirements to Firm the Output of Individual and Aggregated Wind Power Plants," Proc. of the 7th International Conf. on Energy Sustainability and 11th Fuel Cell Science, July 2014, Minnesota. (Submitted).

5. M. Mohanpurkar, D. Zimmerle, S. Suryanarayanan, "An Algorithmic Approach to Detecting Closed Loops in a Power Systems Network," (Under internal review)

Program of Study:

ECE 680A2

ECE 530

ECE 568

ECE 565

ECE 695

ECE 612

PSY 692A-001

N/A