Refine Your Search




Search Results

Technical Paper

Neural Network-Based Diesel Engine Emissions Prediction Using In-Cylinder Combustion Pressure

This paper explores the feasibility of using in-cylinder pressure-based variables to predict gaseous exhaust emissions levels from a Navistar T444 direct injection diesel engine through the use of neural networks. The networks were trained using in-cylinder pressure derived variables generated at steady state conditions over a wide speed and load test matrix. The networks were then validated on previously “unseen” real-time data obtained from the Federal Test Procedure cycle through the use of a high speed digital signal processor data acquisition system. Once fully trained, the DSP-based system developed in this work allows the real-time prediction of NOX and CO2 emissions from this engine on a cycle-by-cycle basis without requiring emissions measurement.
Technical Paper

Artificial Neural Network Based Energy Storage System Modeling for Hybrid Electric Vehicles

The modeling of the energy storage system (ESS) of a Hybrid Electric Vehicle (HEV) poses a considerable challenge. The problem is not amenable to physical modeling without simplifying assumptions that compromise the accuracy of such models. An alternative is to build conventional empirical models. Such models, however, are time-consuming to build and are data-intensive. In this paper, we demonstrate the application of an artificial neural network (ANN) to modeling the ESS. The model maps the system's state-of-charge (SOC) and the vehicle's power requirement to the bus voltage and current. We show that ANN models can accurately capture the complex, non-linear correlations accurately. Further, we propose and deploy our new technique, Smart Select, for designing ANN training data.
Technical Paper

Accurate Models for Complex Vehicle Components using Empirical Methods

Conventional computer models for complex vehicle components (like bushings and dampers) are often inadequate to represent behavior over wide frequency ranges, and/or at large amplitudes. New modeling methods circumvent these limitations by using laboratory measurements with neural networks. The new methods enable accurate simulation for nonlinear, frequency dependent components, having multiple inputs and outputs, under arbitrary excitation. This paper describes one such method, known as Empirical Dynamics Modeling. Examples are presented for vehicle shock absorbers and a rubber bushing. Benefits and limitations are discussed, along with requirements for interfacing to a conventional virtual prototyping environment.
Technical Paper

Development of Shift Control Algorithm Using Estimated Turbine Torque

The powertrain of an automatic transmission has a wide operating range in speed, torque and temperature while driving. It is necessary to know them to achieve good shift quality in various operating conditions without tuning the parameters of the shift quality controller. All but the torque sensor is installed in the automatic transmission because of its high cost. In this study, a more precise algorithm is suggested for estimating turbine torque using a neural network model that has three inputs, i.e., engine speed, turbine speed and temperature. The performance of the suggested turbine torque estimation algorithm is validated through experimental results. To utilize the estimated turbine torque in shift control, a shift control algorithm, which shows good shift quality in various operating conditions, is developed.
Technical Paper

Feasibility Study Of Neural Network Approach In Engine Management System In S.I. Engine

As neural network approach has shown very encouraging results in different fields of engineering. The present work is an attempt to use the NN approach in engine management system to control the air fuel ratio and ignition timing with highest possible accuracy to meet the more and more stringent emission regulation. A feed-forward Neural network with one hidden layer has been used to predict the pulse width & ignition timing. Optimum number of hidden neurons and learning rate were observed to be 13 and 0.7 respectively. After training, the network was validated for 180 sample data and further cross-validated for about 400 data samples. The neural network output results show that the maximum absolute error for pulse width is 0.016 during validation and 0.050 during cross-validation.
Technical Paper

Hybrid Air/Fuel Ratio Control Using the Adaptive Estimation and Neural Network

The paper describes a hybrid air-fuel mixture control system that uses neural network and the direct adaptive algorithm. The A/F ratio stabilization to the stoichiometric value is required to obtain maximum efficiency of the three-way catalytic converter operation. The issues of the algorithm synthesis of the adaptive control of the fuel injection have been formulated. This was supplemented by the presentation of the state-of-the-art in the adaptive control theory as applied to non-stationary random object identification. The control algorithms of the fuel injection have been reviewed and classified. The fuel injection algorithms in the SI engine have been described and differentiated in terms of the used engine model and regulator structure. The algorithms comprise elements of the object modeling as well as adaptive coefficients for the control quality of the air-fuel ratio in the steady and non-steady conditions.
Technical Paper

Estimating the Air/Fuel Ratio from Gaussian Parameterizations of the Ionization Currents in Internal Combustion SI Engines

In this paper we use the idea of parameterizing the ionization current using the sum of two Gaussian functions in an indirect scheme to estimate the AFR. In the first step of the scheme, the Gaussian functions are fitted to the ion signal using a standard least-squares fit. Then, as a second step, the AFR is estimated using the six parameters of the Gaussian functions plus the ignition angle and measurements of the engine speed. The experimental tests show that it is possible to estimate the AFR with good accuracy, using this approach. The best results were obtained using a neural network approach and it is shown in the paper that the AFR can be estimated from the ionization current down to approximately 0.1% in mean square error.
Technical Paper

On Line Working Neural Estimator of SI Engines Operational Parameters

In this paper the evaluation of the suitability of the Artificial Neural Networks for setting up simulation modules for “analytical redundancy” was further carried out. The performance of the ANN modules was enhanced, by taking into account the engine dynamics for the simulation of fast engine transients and obtaining satisfactory results. Working toward actual on board application in Fault Diagnosis systems, some ANN modules were implemented in an on-line system which acquires signals from an engine mounted on a test bench and compares in real time the experimental values with the estimated ones. In this way, it was possible to perform long duration tests of ANN's behaviour, substantially confirming the results of the conventional off-line analysis.
Technical Paper

A Computer Code for S.I. Engine Control and Powertrain Simulation

A computer code oriented to S.I. engine control and powertrain simulation is presented. The model, developed in Matlab-Simulink® environment, predicts engine and driveline states, taking into account the dynamics of air and fuel flows into the intake manifold and the transient response of crankshaft, transmission gearing and vehicle. The model, derived from the code O.D.E.C.S. for the optimal design of engine control strategies now in use at Magneti Marelli, is suitable both for simulation analysis and to achieve optimal engine control strategies for minimum consumption with constraints on exhaust emissions and driveability via mathematical programming techniques. The model is structured as an object oriented modular framework and has been tested for simulating powertrain system and control performance with respect to any given transient and control strategy.
Technical Paper

Closed-Loop Control of Spark Advance and Air-Fuel Ratio in SI Engines Using Cylinder Pressure

The introduction of inexpensive cylinder pressure sensors provides new opportunities for precise engine control. This paper presents a control strategy of spark advance and air-fuel ratio based upon cylinder pressure for spark ignition engines. In order to extend the cylinder pressure based engine control to a wide range of engine speeds, the appropriate choice of control parameters is important as well as essential. For this control scheme, peak pressure and its location for each cylinder during every engine cycle are the major parameters for controlling the air-fuel ratio and spark timing. However, the conventional method requires the measurement of cylinder pressure at every crank angle degree to determine the peak pressure and its location. In this study, the peak pressure and its location were estimated, using a multi-layer feedforward neural network, which needs only five cylinder pressure samples at -40°, -20°, 0°, 20°, and 40° after TDC.
Technical Paper

Approximation and Control of the Engine Torque Using Neural Networks

This paper describes the approximation of the engine torque of an SI-engine using recurrent neural networks. As modern engine control units today are based on engine torque management, there is a need for an accurate determination of the produced engine torque. Since direct measurements of this value are not possible in series applications due to resulting high costs, a fast and accurate algorithm for the approximation of these values from other sensor signals has to be found. Dynamic neural networks are a promising method to fulfill these requirements.
Technical Paper

Utilization of Artificial Neural Networks in the Control, Identification and Condition Monitoring of Hydraulic Systems - An Overview

There has been considerable interest and activity in the area of application of the artificial neural network (ANN) to hydraulic systems. The pattern recognition capabilities of the ANN has led to an early investigation in areas where the neural networks could be trained using signals that were at least statistically similar to those signals which the trained ANN would be exposed during operation. The dynamic and encompassing nature of hydraulic system signals poses more of a challenge to ANN training and implementation than one of only pattern recognition. However, in the past decade, there has been considerable activity and progress in the application of ANN techniques for hydraulic systems control, identification and condition monitoring. This paper provides an overview of work in this area. The ANN has proven to be a valuable addition to the current existing techniques.
Technical Paper

An Intelligent Decision Support System (IDSS) Prototype for Aviation Safety Analysis

An Intelligent Decision Support System (IDSS) integrates Artificial Intelligence (AI) techniques, such as expert systems and neural networks, with classical decision analytic approaches to provide advanced information technology support. This paper discusses the development of an IDSS prototype, termed the Intelligent Safety Performance, Evaluation, and Control (InSPEC) System, that is designed to provide advanced decision support for aviation safety analysis. The five decision support modules of the InSPEC System are briefly described.
Technical Paper

Pyrolysis Processing for Solid Waste Resource Recovery in Space

The NASA objective of expanding the human experience into the far reaches of space will require the development of regenerable life support systems. A key element of these systems is a means for solid waste resource recovery. The objective of this work was to demonstrate the feasibility of pyrolysis processing as a method for the conversion of solid waste materials in a Controlled Ecological Life Support System (CELSS). A pyrolysis process will be useful to NASA in at least four respects: 1) it can be used as a pretreatment for a combustion process; 2) it can be used as a more efficient means of utilizing oxygen and recycling carbon and nitrogen; 3) it can be used to supply fuel gases to fuel cells for power generation; 4) it can be used as the basis for the production of chemicals and materials in space. A composite mixture was made consisting of 10% polyethylene, 15% urea, 25% cellulose, 25% wheat straw, 20% Gerepon TC-42 (space soap) and 5% methionine.
Technical Paper

Optimization of Non-noble Metal based Catalytic Converter for Two-stroke, Two-wheeler Applications and its Performance Prediction

Two wheelers constitute almost three fourth of the vehicular population in developing countries like India, and consequently they are the major contributors of vehicular pollution. Perovskite based catalytic converter is a good cost-effective option except the limitations with perovskite viz., lower surface area, delayed light-off and lower activity. The present paper describes the development of thermally stable and modified alumina washcoat on metallic substrate, suitable for synthesis of perovskite type catalyst. The metallic substrate based catalytic converters with non-noble metal catalyst have been prepared and tested for mass emission conversion efficiency using 2-stroke, 2-wheeler. The results reveal compliance of Indian Emission Norms 2000. The average pressure drop observed was within tolerable limits.
Technical Paper

Modeling of Photosynthesis in Soybean Crops Using Artificial Neural Networks

Important to NASA’s Advanced Life Support program is the development of an autonomous, dynamic, self-contained bioregenerative life support system for future, long duration spacecraft and space stations to provide fresh food, air, water and to recycle waste products. These systems will rely on plants to rejuvenate the air and produce food through the process of photosynthesis and purify water through the process of transpiration. An intelligent, autonomous, reliable, and robust control system must be developed and applied to dynamically manage, control and optimize plant-based life support functions to allow the efficient growth of plants, providing the maximum amount of life essentials while using minimal resources. System identification and modeling of plant growth behavior must first be developed to characterize plant growth functions in order to develop an efficient control system.
Technical Paper

Characterisation of DISI Emissions and Fuel Economy in Homogeneous and Stratified Charge Modes of Operation

An experimental study of the performance of a reverse tumble, DISI engine is reported. Specific fuel consumption and engine-out emissions have been investigated for both homogeneous and stratified modes of fuel injection. Trends in performance with varying AFR, EGR, spark and injection timings have been explored. It is shown that neural networks can be trained to describe these trends accurately for even the most complex case of stratified charge operation with exhaust gas recirculation.
Technical Paper

Emissions Modeling of Heavy-Duty Conventional and Hybrid Electric Vehicles

Today's computer-based vehicle operation simulators use engine speed, engine torque, and lookup tables to predict emissions during a driving simulation [1]. This approach is used primarily for light and medium-duty vehicles, with large discrepancies inherently due to the lack of transient engine emissions data and inaccurate emissions prediction methods [2]. West Virginia University (WVU) has developed an artificial neural network (ANN) based emissions model for incorporation into the ADvanced VehIcle SimulatOR (ADVISOR) software package developed by the National Renewable Energy Laboratory (NREL). Transient engine dynamometer tests were conducted to obtain training data for the ANN. The ANN was trained to predict carbon dioxide (CO2) and oxides of nitrogen (NOx) emissions based on engine speed, torque, and their representative first and second derivatives over various time ranges.
Technical Paper

A Prototype Pyrolyzer for Solid Waste Resource Recovery in Space

Pyrolysis processing is one of several options for solid waste resource recovery in space. It has the advantage of being relatively simple and adaptable to a wide variety of feedstocks and it can produce several usable products from typical waste streams. The objective of this study is to produce a prototype mixed solid waste pyrolyzer for spacecraft applications. A two-stage reactor system was developed which can process about 1 kg of waste per cycle. The reactor includes a pyrolysis chamber where the waste is heated to temperatures above 600°C for primary pyrolysis. The volatile products (liquids, gases) are transported by a N2 purge gas to a second chamber which contains a catalyst bed for cracking the tars at temperatures of about 1000 °C −1100 °C. The tars are cracked into carbon and additional gases. Most of the carbon is subsequently gasified by oxygenated volatiles (CO2, H2O) from the first stage.
Technical Paper


In motor vehicle crashes where an occupant has been seriously or fatally injured from a deploying air bag, a common finding has been that the occupant was in close proximity to the air bag (or out-of-position) at the time of deployment. The occupant may have been out-of-position for a variety of reasons including: driver loss of consciousness, pre-impact braking, multiple impacts, rear facing child seat installation, or late firing of the air bag after the occupant has already been forced against the air bag by the crash deceleration. Considerable research has been initiated to develop new or enhanced injury countermeasures to mitigate injuries to persons, particularly children, who are out-of-position at the time of air bag deployment. This paper reports on the development of an occupant position sensor that might be used in conjunction with dual stage or multi-stage inflation technologies for modulating air bag deployment.