A novel aero-engine control method based on deep reinforcement learning (DRL) is proposed to improve the engine response ability. The Q-learning that is model free and can be performed online is adopted. For improving the learning capacity of DRL, the online sliding window deep neural network (OL-SW-DNN) is proposed and adopted to estimate the action value function. The OL-SW-DNN selects the nearest point data with certain length as training data and is insensitivity to the noise. Finally, the comparison simulations of the proposed method with the proportion-integration-differentiation (PID) that is the most commonly used as an engine controller algorithm in industry are conducted to verify the validity of the proposed method. The results show that, compared with the PID, the acceleration time of the proposed method decreased by 1.525 s under the premise of satisfying all engine limits. INDEX TERMS Aero-engine control method, response ability, deep reinforcement learning, on line, deep neural network.
A new optimization control method, which is based on the optimization employing Bezier curves of control variables, is proposed to improve engine thrust response performance. The minimum of the time consumption for control variables changing from start to end point is selected as objective function. Unlike conventional methods, the optimization variables are altered to the Bezier curves of the control variables in this scheme. This new acceleration optimization control only establishes one optimization problem for the overall acceleration process instead of establishing a series of sub optimization problems, which is usually adopted by conventional optimization controls. Hence, the proposed method is easier to overcome the disadvantage of the conventional ones, which just search for the optimal engine performance in local time period. Finally, for comparisons, the simulations for the proposed method and the conventional one have been both conducted. The simulations demonstrate that, although the conventional optimization control shows better performance in initial half part of acceleration process, however, the overall acceleration performance turns to be worse whereas for the new one, the acceleration time from idle to 95% of the maximum power is surprising decreased by 1.4 s compared to the conventional optimization control.
A real-time optimization control method is proposed to extend turbo-fan engine service life. This real-time optimization control is based on an on-board engine mode, which is devised by a MRR-LSSVR (multi-input multi-output recursive reduced least squares support vector regression method). To solve the optimization problem, a FSQP (feasible sequential quadratic programming) algorithm is utilized. The thermal mechanical fatigue is taken into account during the optimization process. Furthermore, to describe the engine life decaying, a thermal mechanical fatigue model of engine acceleration process is established. The optimization objective function not only contains the sub-item which can get fast response of the engine, but also concludes the sub-item of the total mechanical strain range which has positive relationship to engine fatigue life. Finally, the simulations of the conventional optimization control which just consider engine acceleration performance or the proposed optimization method have been conducted. The simulations demonstrate that the time of the two control methods from idle to 99.5 % of the maximum power are equal. However, the engine life using the proposed optimization method could be surprisingly increased by 36.17 % compared with that using conventional optimization control.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.