Session: 05-04 Fault Detection, Optimization & Uncertainty

Paper Number: 127289

127289 - A Parallel Online Performance Optimization Method for ATR Engines Based on Ensemble Neural Network 

To address the limitations of conventional performance optimization methods for air-turbo rocket (ATR) engines, which are hindered by the accuracy and real-time capabilities of onboard models and thus struggle to yield enhanced optimization outcomes, a novel Ensemble Neural Networks based Parallel Performance Optimization (ENNPPO) methodology is proposed. This strategy amalgamates several small neural networks into a comprehensive larger model for the construction of an advanced onboard ATR engine model, moreover, the independence of each small neural networks allows for the reduction of computational time through parallel processing on GPUs. The method then harnesses the Particle Swarm Optimization (PSO) algorithm to refine performance.  In addition, this paper sets the fluctuation range of the random term in the standard PSO algorithm to decrease with the number of iterations, thus avoiding the fluctuations near the local optimum value in the later stages of the PSO algorithm update.Numerical simulations carried out with embedded GPU computing chips demonstrate that the average optimization time per test point is 0.28 seconds, with an energy consumption of 1.22 Joules. The proposed method achieves an optimization precision that is 78.40% and 84.30% higher than that of traditional neural network approaches and Linear Parameter-Varying (LPV) model methods, respectively. Furthermore, the computation speed is enhanced by a factor of at least 7.8 times than CPU, directly attributable to the expedited parallel processing enabled by GPUs.

Presenting Author: WEIDONG CAI Institute of Engineering Thermophysics, National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine; Chinese Academy of Sciences; School of aeronautics and astronautics at University of Chinese Academy of Sciences

Presenting Author Biography: Weidong Cai received the M.S. degrees in process equipment and control engineering from Beijing University of Chemical Technology, Beijing, in 2021. He is currently pursuing the Ph.D. degree in Mechanical Engineering and Power Machinery at School of aeronautics and astronautics, University of Chinese Academy of Sciences, Beijing, China. His research interests include aero-engine control, machine learning and algorithm acceleration.

Authors:

WEIDONG CAI Institute of Engineering Thermophysics, National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine; Chinese Academy of Sciences; School of aeronautics and astronautics at University of Chinese Academy of Sciences
Wei Zhao Institute of Engineering Thermophysics, Chinese Academy of Sciences; National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine; School of aeronautics and astronautics at University of Chinese Academy of Sciences
Binbin Liu Institute of Engineering Thermophysics, Chinese Academy of Sciences; National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine
Xuesen Yang Institute of Engineering Thermophysics, Chinese Academy of Sciences; National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine
Weiwei Luo Institute of Engineering Thermophysics, Chinese Academy of Sciences; National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine
Qingjun Zhao Institute of Engineering Thermophysics, Chinese Academy of Sciences; National Key Laboratory of Science and Technology on Advanced Light-duty Gas-turbine; School of aeronautics and astronautics at University of Chinese Academy of Sciences; Beijing Key Laboratory of Distributed Combined Cooling Heating and Power System