Search Results

In this paper we show the benefits of integrating a sophisticated engine model into an engine software development process. The core goal is the simulation based tuning of engine control parameters. The work reported here is resulting out of a prolonged cooperation between Siemens VDO Automotive AG and the Institute of Industrial Information Technology, University of Karlsruhe (TH), Germany. The approach is based on a model of the variable energy conversion process within a Diesel engine. The model features phenomenological fuel spray and vaporization models as well as cylinder individual mechanical aspects and fully copes with multiple injection systems. To be useful for an industrial function development process it provides a flexible and modular structure and features computational efficiency - considering real-time capability. The model is matched with the behavior of an engine of interest and connected with a control function under development.

Engine models are a basis for better controlling combustion process and the exhaust emissions resulting from it. Currently the zero- and quasi-dimensional models are mostly used. These types of models are also addressed in this article. Zero-dimensional models are computationally efficient and show good simulation results concerning the in-cylinder pressure. However, by neglecting multi-zonal resolution they are not able to describe fuel efficiency or the generation of pollutants. It is therefore necessary to enhance combustion process models with phenomenological fuel spray and vaporization models, with a local resolution of at least two zones. The chemical model for the calculating of emissions shall be based on a two-zone model. The amount of mass, which is transferred from the unburned to the burned zone, is entered into a chemical model based on the chemical equilibrium for the OCH-system (oxygen/carbon/hydrogen).

The demand for more security, economy, and comfort as well as for a reduced environmental impact increases the importance of electronic components for vehicles. The development of such systems is determined by the requirement of an improved functionality and co-requisite the demand for limited costs. In order to fulfil these demands and taking into consideration the increase of complexity and the melting together to a car wide web, Bosch is developing a structuring concept called CARTRONIC®. This concept is supposed to be open and neutral regarding automotive manufactures and suppliers. The analysis of vehicle control systems via this method is based on formal rules for structuring and modelling. The function-related aspect of CARTRONIC® was represented already at the SAE'98 World Congress. Furthermore the safety-related feature was introduced in more detail at the SAE'99 World Congress. The result of the analysis is an object structure of logical components with defined interfaces.

This article gives an overview of the CARTRONIC® based safety analysis (CSA) including an approach for the automatic determination of failure dependencies in automotive systems. CSA is a safety analysis in an early stage of product development. The goals are to identify safety critical components as soon as practicable in the product development process and to automate the analysis as far as possible. This implies that the system view is abstract, i.e. independent of a certain realization just regarding system functionality. In the CSA so called global failure effects will be systematically identified and assessed regarding severity of potential injuries. Global failure effects are especially important because they reveal failures within the system to the outside world (see also definition 3.1). Additionally the CSA keeps track of failure dependencies and supports the integration of safety measures in the system structure.

The permanent growing traffic volume within the last years leads to novel characteristics of the occuring traffic conditions. However, stochastic prognosis systems on highway corridors still mainly rely on Markovian and ARMA processes [2], possessing the disadvantage of underestimating resulting peaks due to its finite memory. This fact often leads to inadequate prediction measurements and unexpected congestions. Contrary to conventional systems, fractional stochastic algorithms have proved to be a superior alternative in various fields as internet traffic [5], stock market prediction etc. in regard to appropriate peak modeling in the overall traffic situation as well as in microscopic particular sectors [4] (e.g. in order to depict platoons driving along with different speeds). Thus, this paper will first give a clear insight into the theory of fractional modeling.

In the last decade direct injecting diesel engines have become more and more important in the automobile industry. Common-rail systems allow to shape the injection curve to reduce emissions, to get a better fuel economy and to provide more comfort. For this the common-rail system itself must be improved. A fundamental problem of the common-rail technology is to measure the exact injected amount of fuel, depending on the absolute rail pressure and pressure oscillations on the rail. Another problem is the delay between trigger signal of the injector and opening of the injector that results in a non-smooth idle speed. One idea to solve this problem is to install a pressure sensor in front of the injector. Due to costs and density problems under high pressure, a simple, non-invasive sensor is preferred. The sensor introduced in this paper is based on the magneto-elastic effect and does not involve any mechanical elements.

In modern Common Rail injection systems five injections per combustion cycle are state of the art and this number trends to increase. Due to the fast opening and closing of the injectors pressure waves are excited on the pipe between rail and injector. These pressure waves influence the amount of injected fuel mass of subsequent injections. A non-invasive pressure sensor is introduced which allows to measure the pressure on the pipe with little effort and costs. With this sensor it is possible to measure the frequency and, with some limitations, the amplitude of the oscillations. Based on the information delivered by this sensor fuel mass deviations caused by the pressure waves may be compensated in series production vehicles.

The individual development process for distributed, communicating electronic control units hinders the integration of Automotive systems and increases the overall costs. In order to facilitate such applications, services and protocols for Communication, Network Management, and Operating System must be standardized. The aim of the OSEK project is to work out a respective specification proposal in cooperation with several car manufacturers and suppliers. This will permit a cost-effective system integration and support the portation of system functions between different electronic control units.

Modern Diesel engines with direct injection generate high engine torque already at low engine speed. This may cause torsion in the crankshaft of the vehicle which results in oscillations of the whole driveline. The objective of this article is to present a robust controller design concept that takes care of the parametric uncertainty present in the plant and prevents the driveline from oscillating, so-called Anti-Jerk Control. The parameters of the state space model of the drivetrain are identified by measured data. As a measure of oscillations in the driveline the difference between engine speed and wheel speed is used. A H∞ controller is designed using loop shaping and mixed sensitivity approach. The controller concept is evaluated on a test car with a diesel engine. Test results are presented.