Multi Objective Optimization and Mechanism Research of AI Empowered Oil Injection Twin-screw Compressor

Objective: This research focuses on the complex physical processes of oil gas two-phase flow and heat transfer in oil injected twin-screw compressors, and proposes an innovative research paradigm that integrates computational fluid dynamics (CFD) and artificial intelligence. Methods: By establishing a high fidelity CFD model, the system revealed the mechanism of the influence of fuel injection parameters on compressor performance. On this basis, a proxy model based on artificial neural network and Gaussian process regression was constructed to achieve second level prediction of compressor performance. Further combining Bayesian optimization and non dominated sorting genetic algorithm, single objective and multi-objective intelligent optimization with isothermal efficiency, power consumption, and exhaust temperature as objectives were carried out. Results: The results showed that the optimized fuel injection system's isothermal efficiency increased from the benchmark value of 0.758 to 0.801, power consumption decreased by 4.8%, and exhaust temperature decreased by 7.0%. Multi objective optimization has obtained a well distributed Pareto front, revealing the trade-off relationship between performance indicators. Conclusion: This study breaks through the limitations of traditional methods in terms of computational cost and optimization efficiency, providing a new path for the high-end and intelligent development of fluid machinery. The integration of CFD with AI-driven surrogate modeling and multi-objective optimization provides a robust and efficient framework for the performance enhancement of oil-injected twin-screw compressors, offering both mechanistic insights and practical optimal designs for industrial applications.