Models for load forecasting and demand response

Title

Models for load forecasting and demand response

Creator

Margaret, Vijaya.

Contributor

Rao, K Uma

Description

Increasing pressure on the utilities to accommodate energy efficiency, load management and progress in advanced technology has led to transformations of existing grids into smarter grids. With the development of Smart Grid Technology and the integration of smart meters it is possible to control the equipment installed at the consumer site. Creating awareness among the end- users to participate in load management programs instead of capacity addition is the best solution for maintaining the stability in the grid. Utilities can also encourage consumer participation in load control activities. They can ensure that power is given to a consumer during his priority time. For this, loads have to be categorized, prioritized and then considered for load shedding so that revenue loss and social impacts of load shedding are minimized. It would be beneficial if a consumer's load is not completely shed during load shedding. Amount of power that is shed from a consumer can be limited and consumers can be allowed to adjust their loads based on the availability of power and get incentives from the utilities for their change in load pattern. Consumers are also benefited with the reduced energy charges on the consumed energy during these periods. Review of the recent research work shows that demand response and load forecasting play an important role to relieve the power system from economic and environmental constraints. Various approaches have been used in the past for developing different demand response and forecasting methodologies including neural networks, fuzzy logic and statistical techniques. These methodologies fluctuate in complication, suppleness, and information necessity. In addition, statistical methods such as time series, regression, and state space methods have large numerical deviation in the predicted load series. In general, for accurate modeling of nonlinear and undecided type of load behavior, artificial intelligence-based techniques are employed. Also, these methods concentrate mainly on ordinary system conditions. However, proposing the possible Demand Response strategies to maintain power system security constraints in unpredicted turbulences pose a serious challenge. In the undertaken research, a novel load forecasting method using hybrid Genetic Algorithm Support Vector Regression model has been proposed. The forecast error is around 1-2%. The second part of the work focuses on formulation of demand response strategies based on time of the day and load prioritization. A Unique grading method has been proposed to prioritize the loads and load management during power deficiency by controlling the loads individually using different optimization techniques. The performance of three well recognized population based meta-heuristic algorithms such as Genetic Algorithm, Ant Colony Optimization and Particle Swarm Optimization, to solve load management at the consumer level in the Smart Grid environment were examined in terms of their efficiency, effectiveness and consistency in obtaining the optimal solution. In the last part of the work the Demand Response model for residential load is proposed to minimize the energy cost of the electricity usage by shifting the loads from peak period to off-peak period with the help of intelligent techniques such as Artificial Bee Colony Algorithm.

Source

Author's Submission

Date

2018-01-01

Publisher

Christ(Deemed to be University)

Subject

Electrical and Electronics Engineering

Rights

Open Access

Relation

61000096

Format

PDF

Language

English

Type

PhD

Identifier

http://hdl.handle.net/10603/338840