Compound disturbance rejection control of spark ignition–controlled-autoignition hybrid combustion for gasoline engines. (February 2018)

• Record Type: Journal Article
• Authors: Song, Kang; Xie, Hui; Hao, Tianyuan
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 232:Number 2(2018:Feb.)
• Page Start: 264
• Page End: 281
• Publication Date: 2018-02
• Subjects: Spark ignition–controlled-autoignition hybrid combustion -- combustion control -- disturbance rejection control -- extended-state observer -- online identification of the model parameters; Mechanical engineering -- Congresses; Transportation engineering -- Congresses; 629.2
• DOI: 10.1177/0954407017697477
• Languages: English
• ISSNs: 0954-4070

Abstract

Spark ignition–controlled-autoignition hybrid combustion is a promising concept because of its capability to achieve a smooth transition between spark ignition combustion and controlled-autoignition combustion, but it suffers from transient control owing to the high sensitivity to the operating conditions. In this paper, a control solution based on the principle of disturbance rejection is proposed for spark ignition–controlled-autoignition hybrid combustion. The complexity, the non-linearity and the cross-coupling inside are removed by idealizing the combustion process into three independent integrators, for the combustion timing channel, the indicated mean effective pressure channel and the λ (excessive air coefficient) channel respectively. All the other dynamics that deviate from the integrators (internal and external) are 'lumped' together as the total disturbance for each channel. With the total disturbance estimated in real time via the extended-state observer and eliminated by the disturbance rejection law, the enforced plant, i.e. the integrator, is controlled by a simple proportional controller. To enhance the response further, a non-linear model-inversion-based feedforward controller is added. In order to attenuate the slow time-varying disturbances, four correction factors for the model parameters are embedded in the model for online estimation. Validations by both simulations and experiments confirm the superiority of the proposed solution in terms of a fast transient response and a high robustness. By using the bandwidth-parameterization-based extended-state observer tuning method and a Kalman-filter-based extended-state observer, the controller is easy to tune, making it a promising candidate for applications of spark ignition–controlled-autoignition hybrid combustion.