DOI
https://doi.org/10.25772/GYR6-MX08
Author ORCID Identifier
https://orcid.org/0000-0002-4233-2666
Defense Date
2019
Document Type
Thesis
Degree Name
Master of Science
Department
Electrical & Computer Engineering
First Advisor
Zhifang Wang
Abstract
Abstract
Power system load forecasting refers to the study or uses a mathematical method to process
past and future loads systematically, taking into account important system operating
characteristics, capacity expansion decisions, natural conditions, and social impacts, to
meet specific accuracy requirements. Dependence of this, determine the load value at a
specific moment in the future. Improving the level of load forecasting technology is
conducive to the planned power management, which is conducive to rationally arranging
the grid operation mode and unit maintenance plan, and is conducive to formulating
reasonable power supply construction plans and facilitating power improvement, and
improve the economic and social benefits of the system.
At present, there are many methods for load forecasting. The newer algorithms mainly
include the neural network method, time series method, regression analysis method,
support vector machine method, and fuzzy prediction method. However, most of them do
not apply to long-term time-series predictions, and as a result, the prediction accuracy for
long-term power grids does not perform well.
This thesis describes the design of an algorithm that is used to predict the load in a long
time-series. Predict the load is significant and necessary for a dynamic electrical network.
Improved the forecasting algorithm can save a ton of the cost of the load. In this paper, we
propose a load forecasting model using long short-term memory(LSTM). The proposed
implementation of LSTM match with the time-series dataset very well, which can improve
the accuracy of convergence of the training process. We experiment with the difference
time-step to expedites the convergence of the training process. It is found that all cases
achieve significant different forecasting accuracy while forecasting the difference timesteps.
Keywords—Load forecasting, long short-term memory, micro-grid
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
8-12-2019