Use this URL to cite or link to this record in EThOS:
Title: Automotive driveline control by nonparametric QFT methods
Author: Abass, Ahmed Mohamed Fahmy Fathy
ISNI:       0000 0004 2733 6575
Awarding Body: University of Liverpool
Current Institution: University of Liverpool
Date of Award: 2011
Availability of Full Text:
Access from EThOS:
This thesis develops and applies novel identification and control methods for automotive vehicle driveline control. The driveline system has internal backlash combined with resonances which make its control a difficult nonlinear problem. The automotive industry has become very competitive during the last few years and in addition to style and fuel economy, the vehicle performance is a very important issue for the customer. The driveline response to driver demand is a key factor in the customer perception of vehicle quality. A major challenge in its control is that individual driving skills differ greatly among drivers, which leads to a large uncertainty in uncontrolled vehicle response. Current industrial methodologies for the driveline control problem are heuristic and trial and error and require considerable time and resources including extensive on-road vehicle testing. The presented work in thesis in contrast, presents a rapid and systematic methodology which allows the control engineer to achieve pre-determined quantified tracking bounds on the driveline response despite significant system nonlinearity and uncertainty. Firstly a novel mathematical model of the drive line is presented which is subsequently used as one means to evaluate the proposed control methods. Both clutch and backlash nonlinearities are included in the model, and unlike in other published models, backlash is more realistically sandwiched between the compliant clutch and compliant drive shafts. The thesis integrates a Nonparametric (NP) identification approach with Quantitative Feedback Theory (QFT) control methods to result in a novel NP QFT method for nonlinear systems. A NP frequency response identification method is proposed for obtaining a Linear Time Invariant Equivalent (LTIE) sets for the nonlinear system using a frequency weighted windowing method to allow the use of experimentally obtained finite Input/Output (I/O) data records. The NP model is obtained by a local frequency smoothing estimation method in which a frequency set is selected to cover the system bandwidth. This novel NP QFT method has the usual benefits of NP identification such as avoiding concentrating the system information in a limited number of parameters and also permits the acknowledged benefits of QFT control such as the effective linear controller design for nonlinear systems which cannot otherwise be applied without parametric models. A novel technique, based on the discrete Hilbert transform is presented for obtaining an equivalent Minimum Phase (MP) plant for a None Minimum Phase (NMP) nominal plant and determining their phase shift difference which allows the QFT design of such systems based on NP data for the first time. Another application of the NP identification method which is presented is to parameter space control and develops a novel NP parameter space method which has the advantage over QFT of simplicity of controller design and structure at the expense of reduced performance. This method is validated experimentally on the vehicle Internal Combustion (IC) engine idle speed problem. The presented NP QFT method is used to design a controller for a gasoline electronic throttle valve which is a key component of all driveline control systems. Although nonlinear compensation could significantly enhance the outcomes of the process further, it is shown that an effective linear controller can be designed by the NP QFT method without any nonlinear compensation and with an acceptable time response in a quick and systematic method from readily obtained test-data. All used experimental validation approaches use an entirely black box approach in which the controllers are developed directly from experimental testing. The experimental results show that both the new NP parameter space and new NP QFT methods are able to robustly achieve good engine idle speed control and good driveline wheel speed control respectively. In driveline control, the wheel speed response was experimentally found to be always inside the pre-designed boundaries and the controlled system was found able to reject the external disturbances within the desired boundary. The presented techniques can be applied to any similar systems where only experimental test data is available without any need to change the methodology or to use any trial and error sequences. The presented methods provide considerable reduction in the design and testing effort for driveline, electronic throttle valve and idle speed control problems.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available