Multiobjective Reservoir Optimization Using Genetic Algorithm

This thesis investigates the optimization of energy generation in a multipurpose reservoir. The objective functions and the constraints for the multiobjective optimization of reservoir operation for maximization of energy production involve the solution of a set of nonlinear equations governing pressure flow in penstock and turbine system and the hydropower generation. The hydropower generation also requires operational rules that define how the reservoir storage volume must be used at different storage levels. For these reasons, the present thesis uses genetic algorithm (GA) functions available in Matlab to perform the multiobjective optimization of hydropower energy production. Multiobjective optimization is based on two objective functions: maximization of total energy production over a specified number of years for which observed data is available, and maximization of annual firm energy production for individual years. Two separate Matlab codes using different hydropower generation algorithms were written for multiobjective optimization using GA. One of the two Matlab codes disregards the operation of individual turbines when allocating the storage volumes for firm energy and secondary energy and uses a nominal head loss due to the friction. The second Matlab code allocates the storage volumes for firm energy and secondary energy by considering hydropower production by individual turbines and the true head losses for each turbine. In addition to these two Matlab codes for multiobjective GA optimization of the reservoir operation, a third Matlab code was also written to calculate the energy production using a traditional rule-based method. These three Matlab codes were applied to calculate the hydroelectric power generation in a multipurpose reservoir. The reservoir operation strategies determined using GA, with and without the consideration of the operation of individual turbines, were then compared with the those obtained using the rule-based traditional method and the results of the original worksheet analysis provided by the State Water Works (DSI) of Turkey. Using prescribed operational policy and rule-set, this study shows that compared to traditional and DSI results, operations conducted using genetic algorithms produce both higher firm energy and a greater total energy production. Further, these results are found to be accurate over a period of 30 years.

Name thesis:degree_name

M.S. in Engineering Science

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Mechanical Engineering

Year dc:date.available

2018

Author dc:creator

• Tiryakioglu, Seyma

Contributors dc:contributor

• Mustafa S. Altinakar
• Vijay Ramanlingam
• Bahram Alidaee

• Genetic Algorithm
• Optimization
• Reservoir Operation
• Engineering

Repository record dc:identifier

https://egrove.olemiss.edu/etd/435

OAI identifier oai:identifier

oai:egrove.olemiss.edu:etd-1434