【摘要】 针对航空液压泵源测试系统存在的很多非线性特性,在采用矢量控制变频器的基础上,提出神经网络控制器和无源性控制器两种控制策略来控制航空液压泵的转速以测试其特性。基于径向基函数网络和基于对角回归神经网络两种神经网络的控制方案对外界环境因素的影响具有较强的稳定性,后者的控制器的效率更高一些,对非线性有较大的控制潜力,但是其神经网络控制器方案存在不少不足。研究结果表明,当航空液压泵出油口的压力变化时,与神经网络控制器相比,无源性控制器不仅更简单,而且不需要系统参数的精确线性化,实用性强,是一种本质上的非线性控制,航空液压泵在不同压力下的转速具有很好的鲁棒稳定性,能够满足对航空液压泵测试的要求。更多还原

【Abstract】 Aimed at many nonlinear characteristics in the aero hydraulic pump (AHP) test system and on the basis of using vector controlled transducer, two control strategies, i.e. neural network controller (NNC) and passivity based controller (PBC), are proposed to control the rotating speed of AHP for measuring its properties. Both radial basis function (RBF) and diagonal recurrent neural network (DRNN) controllers have strong stabilities to the influence of external environmental factors. And DRNN has higher efficiency and more control potential for nonlinear property but in which there exists many shortages. The results show that, when the pressure changes in the outlet of AHP, the PBC, in which the parameters of a system are not required to be precisely linearized, not only has better practicability but also is an essentially nonlinear control and simpler than the NNC. Under different outlet pressures of AHP, the PBC has good stability and robustness which meet the requirements of testing AHP.更多还原