Applications of fuzzy set theory in reservoir operations

• Main Author: Campbell, Paul Francis
• Format: Others
• Language: English
• Published: 2009
• Online Access: http://hdl.handle.net/2429/4914

Attempts to maximize benefits from hydroelectric reservoirs with mathematical models have been a focus of study for water resource specialists. As increasingly complex models are developed to more closely imitate reality, their usefulness may paradoxically diminish since reservoir operators are less apt to fully understand the models. This results in a general lack of acceptance of mathematical reservoir models amongst the people they were originally developed to serve. Also, the stochastic nature of modelling a system influenced by climatic and economic factors such as a hydroelectric reservoir puts an upper limit on the attainable accuracy of a model. This thesis suggests that a method based on fuzzy set theory may provide a more readily understandable model that recognizes the inherent uncertainties in reservoir modelling. Heuristics or “rules of thumb” are used to simulate operation of a reservoir subject to uncertain inflows and changing hydroelectric power values. This system describes the operation of the reservoir in vague terms such as: “IF predicted inflow is medium-low AND reservoir volume is high THEN suggested outflow is ...“. These rules can be obtained directly from an experienced operator; from analysis of historical data; from data generated by a mathematical simulation of the reservoir or any combination of all three. This thesis illustrates the development and use of a simple fuzzy rulebase for a single hydroelectric reservoir. The rulebase is formed from data generated by a mathematical optimization model (dynamic programming) of the reservoir that simulates several years of operation with random inputs. === Applied Science, Faculty of === Civil Engineering, Department of === Graduate