Search Results

1 In modern turbocharged engines the power output is strongly connected to the turbocharger speed, through the flow characteristics of the turbocharger. Turbo speed is therefore an important state for the engine operation, but it is usually not measured or controlled directly. Still the control system must ensure that the turbo speed does not exceed its maximum allowed value to prevent damaging the turbocharger. Having access to a turbo speed signal, preferably by a cheap and reliable estimation instead of a sensor, could be beneficial for over speed protection and supervision of the turbocharger. This paper proposes a turbo speed observer that only utilizes the conditions around the compressor and a model for the compressor map. These conditions are either measured or can be more easily estimated from available sensors compared the conditions on the turbine side.

The work extends a methodology, for searching for optimal heat release profiles, by adding complex constraints on states. To find the optimum heat release profile a methodology, that uses available theory and methods, was developed that enables the use of state of the art optimal control software to find the optimum combustion trace for a model. The methodology is here extended to include constraints and the method is then applied to study how sensitive the solution is to different effects such as heat transfer, crevice flow, maximum rate of pressure rise, maximum pressure, knock and NO generation. The Gatowski single zone model is extended to a pseudo two zone model, to get an unburned zone that is used to describe the knocking and a burned zone for NO generation. A modification of the extended Zeldovich mechanism that makes it continuously differentiable, is used for NO generation.

In diesel engines with EGR and VGT, the gas flow dynamics has significant nonlinear effects. This is shown by analyzing DC-gains in different operating points showing that these gains have large variations. To handle these nonlinear effects, a nonlinear state dependent input transformation is investigated. This input transformation is achieved through inversion of the models for EGR-flow and turbine flow. It is shown that the input transformation handles the nonlinear effects and decreases the variations in DC-gains substantially. The input transformation is combined with a new control structure that has a pumping work minimization feature and consists of PID controllers and min/max-selectors for coordinated control of EGR-fraction and oxygen/fuel ratio. The EGR flow and the exhaust manifold pressure are chosen as feedback variables in this structure. Further, the set-points for EGR-fraction and oxygen/fuel ratio are transformed to set-points for the feedback variables.

Three methods for compression ratio estimation based on cylinder pressure traces are developed and evaluated for both motored and fired cycles. Two methods rely upon models of polytropic compression and expansion for the cylinder pressure. It is shown that they give a good estimate of the compression ratio, although the estimates are biased. A method based on a variable projection algorithm with a logarithmic norm of the cylinder pressure, which uses interpolation of polytropic models of the expansion and compression asymptotes, is recommended when computational time is an important issue. For motored cycles it yields the smallest bias and confidence intervals for these two methods. For firing cycles a user-specified weighting factor is needed during the combustion phase, which pays off in a smaller estimation bias but also a higher variance. The third method includes heat transfer, crevice effects, and a commonly used heat release model for firing cycles.

It is important to determine the phasing of a measured cylinder pressure trace and crank angle with high accuracy. The reason is that erroneous determination of the position of TDC is a major error source when calculating properties such as heat release etc. A common way to determine the TDC position is to study motored cycles. Heat transfer makes the task more complicated, since it shifts the position of the maximum pressure away from TDC. In this paper a new method for determining the TDC position is proposed that does not require any additional sensors other than a cylinder pressure sensor and an incremental encoder. The idea is to find a point that the cylinder pressure from a motored cycle is symmetric around, since the volume is close to symmetric on either side of TDC. The new method and four published methods are tested and evaluated. Cylinder pressure data used for comparison are from simulations of a SAAB Variable Compression engine.

Mean value cylinder air charge (CAC) estimation models for control and diagnosis are investigated on turbocharged SI-engines. Two topics are studied; Firstly CAC changes due to fuel enrichment and secondly CAC sensitivity to exhaust manifold pressure changes. The objective is to find a CAC model suitable for control and diagnosis. Measurements show that CAC models based on volumetric efficiency gives up to 10% error during fuel enrichment. The error is caused by the cooling effect that the fuel has as it evaporates and thus increases the charge density. To better describe the CAC during fuel enrichment a simple one parameter model is proposed which reduces the CAC estimation error on experimental data from 10% to 3%. With active wastegate control, the pressure changes in the exhaust manifold influences the CAC. The magnitude of this influence is investigated using sensitivity analysis on an exhaust manifold pressure dependent CAC-model.

Most of todays spark-advance controllers operate in open loop but there are several benefits of using feed-back or adaptive schemes based on variables deduced from the cylinder pressure. A systematic study of how different engine conditions change the deduced variables, at optimal ignition timing, is performed. The analysis is performed using a one-zone heat-release model and varying the model parameters. The deduced variables that are studied are: position of the pressure peak, mass fraction burned levels of 30%, 45%, 50%, and 90%, and the pressure ratio. For MBT timing the position for 45% mass fraction burned changed least under a large variety of changes in burn rate. Cycle-to-cycle variations do not have a significant effect and it suffices to evaluate the mean values for the burn rate parameters. The pressure ratio produces values similar to the mass fraction burned and requires no separate treatment.

Heat release analysis by using a pressure sensor signal is a well recognized technique for evaluation of the combustion event, and also for combustion diagnostics. The analysis includes tuning of several parameters in order to accurately explain measured data. This work presents and investigates a systematic method for estimating parameters in heat release models and minimizing the arbitrary choices. In order for the procedure to be systematic there are also the requirements on the model, that it includes no inherent ambiguities, like over-parameterization with respect to the parameters and to the information contained in the measurements. The fundamental question is which parameters, in the heat release model, that can be identified by using only cylinder pressure data. The parameter estimation is based on established techniques, that constructs a predictor for the model and then minimizes a least-squares objective function of the prediction error.

The main result of this paper is a real-time closed loop demonstration of spark advance control by interpretation of ionization current signals. The advantages of such a system is quantified. The ionization current, obtained by using the spark plug as a sensor, is rich on information, but the signal is also complex. A key step in our method is to use parameterized functions to describe the ionization current [1]. The results are validated on a SAAB 2.3 1, normally aspirated, production engine, showing that the placement of the pressure trace relative to TDC is controlled using only the ionization current for feedback.