The identification and the design of an accurate control are still an active researches of nonlinear systems subject to harmonic disturbances. In this paper, we propose two techniques to deal with nonlinear system identification and control problems. First, a new identification scheme based only on input/output measurements is proposed. In this scheme, an adaptive filter is used to extract the periodic disturbance and the non-perturbed system dynamics from the real output. Then, a multimodel approach is adopted to obtain the models base of the real non-perturbed system. Secondly,the models base is used to develop a multi-model sliding mode control (MM-SMC). Thereafter, the structure of plug-in repetitive control is incorporated into the MM-SMC to ensure good closed loop performances in terms of harmonic disturbance rejection and tracking in the case of nonlinear systems. Simulation results of two numerical examples shows the effectiveness of the proposed identification and control strategies in spite the presence of nonlinearities and harmonic disturbances.

精确控制的辨识和设计仍是非线性系统在谐波干扰下的积极研究。在本文中，我们提出了两种技术来处理非线性系统的辨识和控制问题。首先，提出了一种仅基于输入/输出测量的新识别方案。在该方案中，使用自适应滤波器从实际输出中提取周期性扰动和非扰动系统动力学。然后，采用多模型方法获得了真实的非扰动系统的模型库。其次，利用模型库开发了多模型滑模控制（MM-SMC）。之后，MM-SMC中集成了插入式重复控制的结构，以确保在非线性系统情况下在谐波干扰抑制和跟踪方面具有良好的闭环性能。两个数值示例的仿真结果表明，尽管存在非线性和谐波干扰，所提出的识别和控制策略仍然有效。