A Hybrid Dynamic Programming and Neural Network Approach to Unit Commitment with High Renewable Penetration.

Document Type : Research Studies

Authors

1 Prof, Electrical Engineering Department, Faculty of Engineering Mansoura University, Egypt, She is the head of Electrical Engineering Department

2 Electrical Engineering Department Faculty of Engineering Mansoura University

Abstract

This paper presents a solution of the unit commitment (UC) problem for an electrical grid, which contains conventional sources and renewable energy sources as well as storage units. To ensure economical with the stochastic nature of renewable sources, it is essential to develop an efficient forecasting model for renewable power generation. Forecasting model was built by using a hybrid Markov to forecast solar radiation, while, autoregressive integrated moving average model is used to predict wind speed. UC problem incorporates with forecasting, the proposed formulation aims to minimize total production cost. The total production cost includes the fuel costs, environmental cost, operation and maintenance cost (O&M cost), start-up cost, and shutdown cost. UC formulation is subject to multi-constraints. These constraints are system constraint, thermal unit constraints, renewable sources constraints and storage unit's constraint. Also, reserve coefficient is modified to overcome the variation and error of renewable source forecasting by developing two new reserves; up reserve and down spinning reserve. The unit commitment algorithm is solved by simple, fast, and accurate optimization technique. So, hybrid optimization technique used to solve UC is dynamic programming based on neural network. The proposed hybrid techniquemakes the solution faster and more accurate compared with the other techniques. The system under study in this paper is the standard IEEE 30 bus system, with wind speed and solar radiation data of the city of Florida, USA

Keywords

Main Subjects

Files

Share

How to cite

Statistics