Power Research

V. Pavan Kumar ¹, P. S. Kulkarni ²

Affiliations
1 PG Student, Department of Electrical Engineering, VNIT, Nagpur - 440010, IN
2 Associate Professor, Department of Electrical Engineering, VNIT, Nagpur - 440010, IN

Abstract

This paper presents an approach for solving the short-term thermal unit commitment (UC) problem using a multi-layered Feed-forward Back propagation Neural Network (FF-BPNN). The main focus of the paper is on finding the schedule of committed thermal units within a short computational time such that the total operating cost is minimized. The proposed method is implemented and tested on a 3-unit and 10-unit systems for a scheduling period of 4-hours and 24-hours respectively in MATLAB software using the Neural Network toolbox. Comparison of simulation results of the proposed method with the results of previous published methods shows that the proposed FF-BPNN method provides better solution with less computational time.

Keywords

Artificial Neural Networks (ANN), Dynamic Programming (DP), Multi-layered Feedforward Back propagation Neural Network (FF-BPNN), Lagrangian Relaxation (LR), Priority List (PL), Unit Commitment (UC)

Full Text

     

V. Pavan Kumar ¹, P. S. Kulkarni ²

Affiliations
1 PG Student, Department of Electrical Engineering, VNIT, Nagpur, IN
2 Associate Professor, Department of Electrical Engineering, VNIT, Nagpur, IN

Abstract

This paper presents an approach for solving the short-term thermal unit commitment (UC) problem using a multi-layered Feed-forward Back propagation Neural Network (FF-BPNN). The main focus of the paper is on finding the schedule of committed thermal units within a short computational timesuch that the total operating cost is minimized. The proposed method is implemented and tested on a 3-unit and 10-unit systems for a scheduling period of 4-hours and 24-hours respectively in MATLABTM software using the Neural Network toolbox. Comparison of simulation results of the proposed method with the results of previous published methods shows that the proposed FF-BPNN method provides better solution with less computational time.

Keywords

Artificial Neural Networks (ANN), Dynamic Programming (DP), Multi-layered Feedforward Back propagation Neural Network (FF-BPNN), Lagrangian Relaxation (LR), Priority List (PL), Unit Commitment (UC)

Full Text