S. Yurkovich

On the Control of Engine Start/Stop Dynamics in a Hybrid Electric Vehicle

The starter/alternator technology is considered an easily realizable hybrid electric vehicle (HEV) configuration to achieve significant fuel economy without compromising consumer acceptability. Several examples can be found in production or near-production vehicles, with implementation based on a spark ignition (SI) engine coupled with either a belted starter/alternator (BSA) or an integrated starter/alternator (ISA). One of the many challenges in successfully developing a starter/alternator HEV is to achieve engine start and stop operations with minimum passenger discomfort. This requires control of the electric motor to start and stop the engine quickly and smoothly, without compromising the vehicle noise, vibration, and harshness signature. The issue becomes more critical in the case of diesel hybrids, as the peak compression torque is much larger than in SI engines. This paper documents the results of a research activity focused on the control of the start and stop dynamics of a HEV with a belted starter/alternator. The work was conducted on a production 1.9 l common-rail diesel engine coupled to a 10.6 kW permanent magnet motor. The system is part of a series/parallel HEV powertrain, designed to fit a midsize prototype sport utility vehicle. A preliminary experimental investigation was done to assess the feasibility of the concept and to partially characterize the system. This facilitated the design of a control-oriented nonlinear model of the system dynamics and its validation on the complete HEV hardware. Model-based control techniques were then applied to design a controller for the belted starter/alternator, ensuring quick and smooth engine start operations. The final control design has been implemented on the vehicle. The research outcomes demonstrated that the BSA is effective in starting the diesel engine quickly and with very limited vibration and noise.

Fault Detection Using Set-Membership Identification

Abstract In this paper, two novel approaches for detection of faults in dynamical systems are presented. Both approaches are based on set-membership identification, a system identification strategy which seeks to identify a set of parameters rather than a single point estimate. The optimal volume ellipsoid algorithm (OVE) and the OVE algorithm for time-varying systems (OVETV) will be utilized for set-membership identification. The first detection strategy uses a consistency check which is integral to the OVE and OVETV algorithms. The second approach combines an ellipsoid intersection test with the OVETV algorithm.

Model Predictive Control as an Energy Management Strategy for Hybrid Electric Vehicles

The energy management strategy in a hybrid electric vehicle is viewed as an optimal control problem and is solved using Model Predictve Control (MPC). The method is applied to a series hybrid electric vehicle, using a linearized model in state space formulation and a linear MPC algorithm, based on quadratic programming, to find a feasible suboptimal solution. The significance of the results lies in obtaining a real-time implementable control law. The MPC algorithm is applied using a quasi-static simulator developed in the MATLAB environment. The MPC solution is compared with the dynamic programming solution (offline optimization). The dynamic programming algorithm, which requires the entire driving cycle to be known a-priori, guarantees the optimality and is used here as the benchmark solution. The effect of the parameters of the MPC (length of prediction horizon, type of prediction) is also investigated.Copyright

Parameter set estimation algorithms for time-varying systems

Parameter set estimation (PSE), a class of system identification schemes which aims at characterizing the uncertainty in the identification experiment, is philosophically different from traditional parameter estimation schemes which seek to identify a single point (model) in the parameter space. The literature has seen a good deal of attention paid to PSE techniques in recent years, primarily because it is projected that they will play a vital role in robust identification for control. An important step in current research along these lines is development of PSE algorithms for systems which are time varying in nature; this is particularly true if the identified model set is to be used in an adaptive setting, such as for gain scheduling or autotuning. In this paper, we extend an ellipsoid algorithm for PSE of time-invariant systems to time-varying systems. We show how knowledge of dependences in the parameter variations can be exploited to reduce the number of computations in the resulting algorithm. Finally, scalar bound inflation, a second strategy for PSE of time-varying systems, is optimized for volume, and a comparison of the two algorithms is made.

Feedback Linearization with Acceleration Feedback for a Two-Lin Flexible Manipulator

This paper utilizes a three-loop control strategy in a decentralized fashion for each link of a two-link flexible manipulator. The first loop employs techniques from feedback linearization for the rigid dynamics, where motor dynamics are accounted for in the model and the linearizing control law. The second (outer) loop consists of a simple proportional-derivative (PD) control law for accurate rigid body angle tracking. The third outer loop employs acceleration feedback, from sensed linear acceleration at each link endpoint, to account for flexure effects. The overall scheme is relatively simple in order to facilitate easy implementation; experimental results are provided to verify the effectiveness of the developed schemes.

Multi-variable control for an automotive traction PEM fuel cell system

This paper describes an approach for the control of pressurized PEM fuel cell systems used in automotive traction applications. This model-based controller design approach is based on a 13-states non-linear dynamic model. The focus of the paper is in controlling the excess air ratio while tracking an optimum variable pressurization for maximum system efficiency during load transients. The control approach combines a feed-forward approach based on the steady state plant inverse response, coupled to a multi-variable LQR feedback controller. The controller shows excellent performance over severe load transients with both actual states and state observers feedback

Decentralized Frequency Shaping and Modal Sensitivities for Optimal Control of Large Space Structures

Abstract In recent years methods from the area of decentralized control have emerged for analysis and control of large flexible space structures. Many critical issues remain for consideration in control problems of flexible spacecraft, including the need for incorporation of actuator dynamics in the system model, the need for an initial stabilizing feedback solution to initiate computation in optimal controller design, and the need for inclusion of frequency domain constraints into the state-space formulation. In this paper we present approaches cast in a decentralized setting which address each of these issues.

Vibration Control for Slewing Flexible Structures

In this paper we develop several different controllers for slewing flexible structures with results evaluated on a flexible planar truss testbed. The objective of these control schemes is rapid large angle maneuvering (slewing) of the truss with effective suppression of residual vibrations, particularly at the endpoints. Results of three control schemes are presented, two linear schemes based on input shaping and one non-linear scheme. The second of the linear controllers is a novel approach to incorporate a discrete-time input shaper into the feedback loop. The nonlinear scheme consists of an implementation of sliding mode control.

Application of extended Kalman filter to on-line diesel engine cylinder pressure estimation

Cylinder pressure is one of the most important parameters characterizing the combustion process in an internal combustion engine. The recent developments in piezoelectric pressure transducers and progress in on-line computational throughput facilitate the use of cylinder pressure as a feedback signal for engine combustion control. However, a typical production cylinder pressure sensor is subject to noise and offset issues that require signal processing methods, including averaging over several engine cycles, in order to extract a pressure trace sufficiently accurate for combustion characterization. This limits the application of cylinder pressure sensing to off-line applications. In order to enable closed-loop combustion control using cylinder pressure feedback, this study proposes a real-time estimation algorithm that extracts the pressure signal on a crank-angle basis. A simplified thermodynamic model for Diesel engine combustion is derived to predict the in-cylinder pressure. The model is then adapted to model-based estimation, by applying an Extended Kalman Filter in conjunction with a recursive least squares estimation. The resulting estimator is tested on a high-fidelity Diesel engine model for different operating conditions. The results obtained show the effectiveness of the estimator in reconstructing the cylinder pressure and in rejecting measurement noise and modeling errors.Copyright

Decentralized variable structure control of a two-arm robotic system

The control problem for a two-arm robotic system in coordinated motion is addressed. A hierarchical framework, employing two levels of control hierarchy, is utilized, and the decentralized model reference adaptive control approach using variable structure controllers (DMRA-VSC) is applied. Within the control hierarchy, the DMRA-VSC strategy is accomplished at the lower level, where control is responsible for the servoing of each joint. These local controllers are coordinated by the high-level, central controller, whose task is to provide the local controllers with the upper bound on the dynamical interactions with other subsystems. The local controllers are responsible for making each link follow the prescribed local reference subsystem in moving from one position to another. This is done using the local measurements and the information provided by the central controller. Advantages of the DMRA-VSC approach for multiple manipulator control include the inherent robustness properties to nonlinearities and interaction effects, the decentralization structure facilitating ease in multiple manipulator system programming and implementation, and the general structure of the controller which allows further extensions such as force feedback.