广义脉冲响应函数模型由于其辨识对输入激励的苛刻要求,限制了其在非线性系统故障诊断方面的应用。提出了一种基于广义脉冲响应函数的可以描述非线性系统动态响应过程的N阶线模型。它将广义脉冲响应函数的各阶空间特征压缩到一条边线上,减少模型的表征量,从而降低辨识的输入激励要求,使被压缩的模型更易于被辨识。并且,根据广义脉冲响应函数的特性,提出了模型的辨识方法。最后,仿真验证了模型的辨识,并分析了故障对模型的影响特征。表明该模型适用于非线性系统的故障诊断。 The general impulse response function model has limited its application in the fault diagnosis of nonlinear systems because of its demanding identification of input stimuli. A N-order model based on generalized impulse response function is proposed to describe the dynamic response of nonlinear systems. It compresses each spatial feature of generalized impulse response function to an edge and reduces the characterization quantity of the model, so as to reduce the input excitation requirement of identification and make the compressed model easier to be identified. And, according to the characteristics of generalized impulse response function, a method of model identification is proposed. Finally, the simulation verifies the model identification and analyzes the influence of the fault on the model. It shows that the model is suitable for the fault diagnosis of nonlinear systems.