Loading...

Developing Prerequisites for the Implementation of Incentive-Based Demand Response – Predicting Demand Response Potential and Estimating the Baseline Load Pattern Using Machined Learning Methods

Naserian, Mehdi | 2022

30 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56060 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Fotuhi Firuzabad, Mahmud
  7. Abstract:
  8. In recent years, the development of demand response programs with the aim of activating the demand side in order to achieve a wide range of economic and technical benefits in the development of the power system structure has been considered. With the implementation of incentive-based demand response plans, customers receive financial rewards based on changing their load pattern in the favor of electricity companies. However, mid-term and long-term planning to take advantage of this potential requires accurate estimation of the demand response potential in the power system. Also, the value of the baseline load curve of the customers (an estimate of the energy consumption of the customers if they do not participate in the demand response program) which will be required to calculate the amounts of financial rewards is unknown. Therefore, in order to calculate the financial reward of customers, it is necessary to estimate the value of the baseline load curve in some way. For this purpose, considering the promising performance of machine learning methods in solving complex problems in similar fields, it is suggested to use them to estimate the demand response potential and the baseline load of customers. In this regard, the aim of this project is to provide the required platform for the implementation of incentive-based demand response plans, including forecasting the demand response potential and estimating the baseline load of customers by using machine learning methods. In this research, 3 methods are presented to predict the demand response potential. In the first method, by using mathematical modeling of flexible household equipment and by providing a new and multi-level reward system and considering the convenience of customers, the demand response potential of this equipment has been estimated. In the second method, the demand response potential of the air conditioning systems is estimated for a large number of customers and the high-consumption customers are separated from the low-consumption ones. In the third method, by knowing the demand response potential of a number of customers, resulting from participation in past demand response programs and the extracted technical and non-technical features, models have been developed for predicting the demand response potential of new or repeated customers in a new demand response program with high accuracy. In order to estimate the baseline load of customers, extensive models have been developed and compared in terms of speed and accuracy
  9. Keywords:
  10. Demand Response ; Machine Learning ; Baseline Load Estimation ; Incentive-Based Program

 Digital Object List

 Bookmark

No TOC