Optimal design techniques applied to small islanded energy systems

摘要

The Thesis explores the optimisation of Small Islanded Energy Systems (SIES). This work defines a small energy systems as one that is designed to meet Electrical and Hot Water loads in the range of 5 to 250 kWhrs per day. The process of optimisation is defined as the process of establishing the design configuration (size of generating arrays, storage and generator size) and operating rules such that the long term Cost of Energy (CoE) together with system CO2 emissions are minimised. The early sections of the thesis reviews existing techniques for energy system optimisation and explores advantages and disadvantages of different techniques. Through this process a question emerged regarding the ability of existing techniques to address the impact of consecutive days of incident solar energy on system designs. A concept is developed that this question of consecutive incident energy days could be addressed by a two stage design optimisation where the first stage is based around optimising the system design for the Modal day of incident energy and then iterating that solution based on the probability of occurrence of a particular incident energy on the days follow- ing the baseline Modal day. A review of suitable mathematical techniques to execute the suggested approach is conducted and a technique from the discipline of Operations Research (OR) referred to as Stochastic Programming with Recourse(SPwR) is nominated as a likely possible methodology The ability of this combination of a two stage optimisation, based around a baseline Modal day of incident energy executed using Stochastic Programming with Recourse together with Probabil- ity Distribution Functions of incident solar energy to produce optimal solutions is explored using a range of increasingly complex systems. The work concludes that the technique can be used to opti- mise the design of small energy systems to meet both electrical and hot water loads. The technique developed allows pre-processing of weather data accommodates load variation with weather and allows consideration of a range of commercial issues associated with grid connection of SIES.