College - Author 1

College of Engineering

Department - Author 1

Electrical Engineering Department

Degree Name - Author 1

BS in Electrical Engineering



Primary Advisor

Xiao-Hua (Helen) Yu, College of Engineering, Electrical Engineering Department


Electricity price depends on numerous factors including the weather, location, time of year/month/day (summer, holidays, day/night, etc.), consumption, and market changes or disruptions. The power loss in the transmission lines and the use of electricity also affect the load's price. Forecasting about the electricity price provides future trends and patterns consumption of the users. Forecasting electricity price is crucially important for producers and consumers in the energy trading markets. It is a complicated task because of the uncertainty behaviors and demand fluctuation. The maximization of profit for participants is highly associated with the bidding strategies. Multiple forecasting electricity price tools were proposed such as using probabilistic forecast, multivariate factor models, or combining forecast, however, none can be generalized for all demand patterns. Therefore, this paper presents a methodology that can outperform others under different application settings in this problem. Long-short term memory (LSTM) method has been used for classifying, processing, and making predictions based on time series data. It was developed to solve the vanishing gradient problem that happens when training with Recurrent Neural Network (RNN). Tensorflow platform are used for training the proposed LSTM network and persistence forecast network model. The hourly price data from New England day-ahead markets are used in this study. The performance of this method is also compared with other existing methods.

Available for download on Sunday, June 08, 2025