Search Results

As a physical description of the emissions of a specific engine is seldom possible, we present here a method to design an online dynamic estimator for PM and NOx based on data. The design method is based on a systematic search of function candidates performed using genetic programming after data have been pre-treated in an adequate fashion. While data and a simple data pretreatment prove enough for NOx, some basic physical understanding is necessary to preset the method and obtain the required precision in the case of PM. The method has been applied for raw emissions of a production DI diesel engine and shows a remarkable prediction performance.

Complexity and nonlinearity of engines makes precise first principle engine models often difficult to obtain, as for instance for emissions. System identification is a well known possible alternative, successfully used in several automotive applications. In most cases system identification is concerned with the estimation of the unknown parameters of a known set of equations. Unfortunately, for many engine subsystems, there is no sufficiently precise or real time suitable model. This paper presents a sequential algorithm which allows to derive real time suitable models on line by a combination of model structure hypothesis of increasing complexity and an associated optimal input design and selection process. This paper introduces the method and shows its use both for a rather simple and a very difficult engine identification task, a dynamical model of the airpath of a Diesel engine and a dynamical model of nitrogen oxides and particulate matter.

Advanced driver assistance systems like cooperative adaptive cruise control (CACC) are designed to exploit information provided by vehicle-to-vehicle (V2V) and/or infrastructure-to-vehicle (I2V) communication systems to achieve desired objectives such as safety, traffic fluidity or fuel economy. In a day to day traffic scenario, the presence of unknown disturbances complicates achieving these objectives. In particular, CACC benefits in terms of fuel economy require the prediction of the behavior of a preceding vehicle during a finite time horizon. This paper suggests an estimation method based on actual and past inter-vehicle distance data as well as on traffic and upcoming traffic lights. This information is used to train a set of nonlinear, autoregressive (NARX) models. Two scenarios are investigated, one of them assumes a V2V communication with the predecessor, the other uses only data acquired by on-board vehicle sensors.

It is well known that driver behavior can affect fuel consumption to a large extent hence modifying it can lead to reasonable reduction in the magnitude of 10 to 20%. However, it is also known that effects of training are short lived and therefore many authors and companies suggest the use of monitoring systems, sometimes called eco-driver, which allow recognizing opportunities for reduction. V2V is an emerging technology which has been widely studied especially for safety applications. In terms of fuel consumption, there has also been a significant effort for methods directed to coordinate the movement of vehicles, especially of trucks, to improve fuel consumption by platooning [1].

Transient emission peaks have become an important fraction of the total emissions during the standardized test cycles for passenger car Diesel engines. This paper is concerned with their reduction, in particular of nitric oxides (NOx) and particulate matter (PM) emissions, by online receding horizon optimal control. It is based on former works in which alternative target quantities for engine control were proposed, namely in-cylinder oxygen concentrations before (O2,BC) and after combustion (O2,AC). The actual work is concerned with testing an in-cylinder oxygen concentrations based control in simulation as well as by a real-time implementation on a turbocharged common rail passenger car production Diesel engine. The promising results confirm the choice of these concentrations as sensible control references and the feasibility of a real-time use in a model predictive control implementation.

On-line measurement of engine NOx emissions is the object of a substantial effort, as it would strongly improve the control of CI engines. Many efforts have been directed towards hardware solutions, in particular to physical sensors, which have already reached a certain degree of maturity. In this paper, we are concerned with an alternative approach, a virtual sensor, which is essentially a software code able to estimate the correct value of an unmeasured variable, thus including in some sense an input/output model of the process. Most virtual sensors are either derived by fitting data to a generic structure (like an artificial neural network, ANN) or by physical principles. In both cases, the quality of the sensor tends to be poor outside the measured values.

Modern Diesel engines have become complex systems with a high number of available sensor information and degrees of freedom in control. Due to recent developments in production type in-cylinder pressure sensors, there is again an upcoming interest for in-cylinder pressure based applications. Besides the standard approaches, like to use it for closed loop combustion control, also estimation and on-board diagnostics have become important topics. Not surprising in general the trend is to utilize those sensors for as many tasks as possible. Consequently this work focuses on the estimation of the injection parameters based on the indicated pressure signal information which can be seen as first step of a combustion control based on desirable indicated pressure characteristics which may be utilized for e.g. the minimization of NOx emissions. Currently the acquisition of the cylinder pressure traces can be done in real-time by fast FPGA (Field Programmable Gate Array) based systems.

Pollutant immissions must be kept below some threshold values to prevent health and environmental damage. At the moment, the problem is usually met by constant emission limits for each vehicle independently from specific conditions - in particular, without any relation to the actual immission situation. This approach offers the advantage of simplicity, but offers no guarantee that the immission levels will be kept. New developments, in particular the expected diffusion of i2v methods, allows suggesting context specific emission levels so that the total emission roughly corresponding to the local immissions - can be limited to the target values. To meet this goal, emission-oriented control will be needed. This paper proposes a robust control system which allows tracking a time-varying NOx profile, based on the sliding mode concept.

This paper shows the potential of a multicalibration approach for reducing fuel consumption while keeping pollutant immissions. The paper demonstrates that the current engine control approach with a single fixed calibration involves important fuel penalties in areas with low vehicle densities where local pollution is not an issue, while the NOx emissions in urban areas are usually too high to fulfill air quality standards. The proposed strategy is based on using information about the vehicle location and the NOx concentrations in the ambient to choose a suitable calibration amongst a set of possibilities. To assess the potential of such a strategy experimental tests have been done with a state-of-art turbocharged Diesel engine. First, a design of experiments is used to obtain three different calibrations.

Especially in view of more and more stringent emission legislation in passenger cars it is required to reduce the amount of pollutants. In the case of Diesel engines mainly NOx and PM are emitted during engine operation. The main influence factors for these pollutants are the in-cylinder oxygen concentration and the injected fuel amount. Typically the engine control task can be divided into two separate main parts, the fuel and the air system. Commonly air system control, consisting of a turbocharger and exhaust gas recirculation control, is used to provide the required amount of oxygen and address the emission targets, whereas the fuel is used to provide the desired torque. Especially in transient maneuvers the different time scales of both systems can lead to emission peaks which are not desired. Against this background in this work instead of the common way to address the air system, the fuel system is considered to reduce emission peaks during transients.

Nitrogen oxide (NOx) sensing is required both for on-board diagnosis and optimal selective catalytic reduction (SCR)-catalyst control in heavy duty diesel engines. This can be accomplished either by physical solid-state sensors, or by so-called virtual sensors, which estimate the value of the target quantity using other states by means of a model. Both approaches have advantages and disadvantages. This paper resumes the derivation and the identification of a virtual sensor based on a polynomial structure and optimal experimental design methods and compares its performance to the one of a production physical solid state sensor. The virtual sensor is compared with a commercially available solid-state sensor in terms of accuracy (stationary as well as dynamic) and operation limits.

The focus in the development of modern exhaust after treatment systems, like the Diesel Oxidation Catalyst (DOC), the Diesel Particulate Filter (DPF) and Selective Catalytic Reduction (SCR), is to increase on one hand the oxidation rates of Carbon monoxide (CO), HC (Hydro Carbons) and NO (Nitrogen Oxide) and on the other hand the reduction rates of Particulate Matter (PM) and the NOx emissions to fulfill the more and more restricting requirements of the exhaust emission legislation. The simplest, practical most relevant way to obtain such a dosing strategy of a SCR system is the use of a nonlinear map, which has to be determined by extensive calibration efforts. This feedforward action has the advantage of not requiring a downstream NOx sensor and can achieve high conversion efficiency under steady-state operating conditions for nominal systems.

Adaptive cruise control (ACC) systems allow a safe and reliable driving by adapting the velocity of the vehicle to velocity setpoints and the distance from preceding vehicles. This substantially reduces the effort of the driver especially in heavy traffic conditions. However, standard ACC systems do not necessarily take in account comfort and fuel efficiency. Recently some work has been done of the latter aspect. This paper extends previous works for CI engines by incorporating a prediction model of the surrounding traffic and a simplified control law capable for real time use in experiments. The prediction model itself uses sinusoidal functions as the traffic measurements often show periodic behavior and is adapted in every sample instant with respect to the predecessor's velocity. Furthermore, the controlled vehicle is forced to stay within a specific inter-vehicle distance corridor to avoid collisions and ensure safe driving.