For their almost zero carbon emission supercritical CO2 cycles are a promising power generation technology that offers numerous benefits in a wide range of applications that include solar power, marine, biomass, waste heat recovery, and next-generation nuclear reactors. The Carleton University Brayton Cycle Loop (CUBCL) project involves almost all aspects of a 10MW S-CO2 gas turbine design, including thermodynamic performance analysis, aerodynamic and structural design, heat exchanger and materials selection, dynamic modeling, control systems design and condition monitoring. This work is to develop a health monitoring system for the 10 MW S-CO2 gas turbine. Taking this technology to a commercial level will require effective condition monitoring. S-CO2 gas turbine power plant has considerable potential to provide sustainable power generation systems that can obtain higher plant efficiency, and one of the most important factors to consider is the purity of the working fluid. Since much of S-CO2 cycle performance improvement is derived from the physical properties of supercritical CO2, the cycle efficiency will decrease as the CO2 concentration decreases. The carbon dioxide near the critical point does not follow the ideal gas law. Due to the extreme variability in properties of S-CO2, the need of the developed monitoring model framework provides the proper 10 MW S-CO2 turbine performance predictions that are required not only during the design stages but also for further cycle specific analysis during the turbine operations. This thesis investigates methodologies for condition monitoring for the 10MW S-CO2 gas turbine using Artificial Neural Networks (ANNs). Results of the study show that the impurities affect all turbine component but recompressor is the most affected by overheating and required further protection. The current design of the 10 MW S-CO2 gas turbine recompressor works on fluid before heat is removed by the pre-cooler. Fouling in the recuperators raises the recompressor inlet temperatures due to the poor heat transfer between two streams which can cause recompressor failure due to overheating. This highlighted the need for an additional protection to prevent the recompressor from overheating. The result shows that the NARX model successfully capture the degradation of the turbine performance due to impurities ....