Three methods of identification of the linear part of electronic throttle body (process) model have been presented. Individual parameters of the physical linear model were estimated by conducting a series of identification experiments. From the stand point of electronic throttle control system design, it is more convenient and accurate to identify black-box linear models. Identification of the integral + lag process model is more convenient for on-line applications (e.g. electronic throttle auto-tuning), while the ARX model identification is generally more accurate, and thus convenient for off-line identification. Identification of nonlinear ETB effects has included identification of different static and dynamic friction effects, as well as identification of static curve nonlinearity at the limp-home position. It has been shown that electronic throttle friction includes significant presliding displacement effect. In order to capture this effect, so-called hybrid dynamic friction model has been proposed and experimentally validated. A nonlinear strategy of electronic throttle control has been presented. The strategy consists of linear feedback and feedforward controllers, and nonlinear friction and limp-home feedback compensators. The presented experimental results have shown efficient rejection of the friction and limp-home effects, which resulted in fast and accurate control system responses for different electronic throttle operating conditions. In order to improve the control system robustness, the control strategy has been extended with an adaptation mechanism consisting of auto-tuning and self-tuning algorithms. The auto-tuner application provides that all electronic throttles from a vehicle production series have similar (desired) control performance, regardless of slow variations of electronic throttle body parameters due to production deviations, external conditions variations, and aging. The auto-tuner does not require any prior knowledge of the process parameters. It is characterized by simple implementation and short execution time (about 1.5 sec). The self-tuning strategy has been developed in order to deal with process parameter variations which occur during a single engine run (i.e. those that cannot be captured by occasional utilization of the auto-tuner). These parameters are the battery voltage, the armature resistance, and the limp-home position. The battery voltage is directly measured, while the armature resistance and limp-home position are estimated on-line based only on the standard electronic throttle measurement signals. The results of experimental verification of the self-tuning strategy have demonstrated fast and accurate process parameter estimation, and consequently superior performance of the adaptive control strategy compared to the classical, non-adaptive strategy.
An adaptive nonlinear strategy of electronic throttle control
2004
13 Seiten, 23 Bilder, 14 Quellen
Conference paper
English
Adaptivregelung , Algorithmus , Drosselregelung , elektronische Regelung , experimentelles Ergebnis , Gleichstrommotor , Kompensator (Elektrotechnik) , Kraftfahrzeugelektronik , lineares System , nichtlineare Regelung , Parameteridentifikation , Parameterschätzung , PID-Regler , Selbstabstimmung , Servoantrieb
An Adaptive Nonlinear Strategy of Electronic Throttle Control
British Library Conference Proceedings | 2004
|An adaptive nonlinear strategy of electronic throttle control
Automotive engineering | 2004
|An Adaptive Nonlinear Strategy of Electronic Throttle Control
SAE Technical Papers | 2004
|2004-01-0897 An Adaptive Nonlinear Strategy of Electronic Throttle Control
British Library Conference Proceedings | 2004
|Adaptive control of automotive electronic throttle
Tema Archive | 2006
|