Abstract
Industrial waste heat, biomass heat, geothermal heat, and solar thermal energy belong to lowtemperature thermal energy resources. They were rarely used for power generation in the past and were completely wasted due to their low power conversion efficiency. However, their utilization can reduce the proportion of thermal power generation in Taiwan if the power generation efficiency can be enlarged. The utilization can even further reduce greenhouse gas emissions and air pollution. Because the temperature of the medium and low temperature heat source, the thermal efficiency of the working elements of the power generation system, and environmental conditions are all different, the commonly used way to improve the power generation efficiency is to connect different types of thermal cycles to improve the power generation efficiency through different working fluids in the cycle. However, it will increase the scope, number and complexity of variable parameters. In order to systematically solve the multi-parameter optimization problem, this study combined the optimization theory with the thermal properties database to write a program. It will optimize the common low-temperature heat source power generation cycle (Kalina cycle) within the known parameters (expander inlet pressure, temperature, working fluid concentration) to improve power generation efficiency. In this study, the above thermal cycle is first established into a numerical thermal model, which reproduces the results of previous literature to verify the accuracy of the thermal model. Then, the model is combined with a self-written genetic algorithm code to optimize the important parameters of thermal model. Finally, a set of optimized parameters is obtained, so that the optimal power generation efficiency obtained by the Karina cycle is 14.652%. The results of this research prove that the genetic algorithm can effectively solve the multi-parameter optimization problem and help design the cycle system to achieve the best efficiency.