J. A. Drallmeier

Time Irreversibility and Comparison of Cyclic-Variability Models

We describe a method for detecting and quantifying time irreversibility in experimental engine data. We apply this method to experimental heat-release measurements from four sparkignited engines under leaning fueling conditions. We demonstrate that the observed behavior is inconsistent with a linear Gaussian random process and is more appropriately described as a noisy nonlinear dynamical process.

Neural Network Control of Spark Ignition Engines with High EGR Levels

Research has shown substantial reductions in the oxides of nitrogen (NOx) concentrations by using 10% to 25% exhaust gas recirculation (EGR) in spark ignition (SI) engines (Dudek and Sain, 1989). However under high EGR levels the engine exhibits strong cyclic dispersion in heat release which may lead to instability and unsatisfactory performance. A suite of neural network (NN)-based output feedback controllers with and without reinforcement learning is developed to control the SI engine at high levels of EGR even when the engine dynamics are unknown by using fuel as the control input. A separate control loop was designed for controlling EGR levels. The neural network controllers consists of three NN: a) ANN observer to estimate the states of the engine such as total fuel and air; b) a second NN for generating virtual input; and c) a third NN for generating actual control input. For reinforcement learning, an additional NN is used as the critic. The stability analysis of the closed loop system is given and the boundedness of all signals is ensured without separation principle. Online training is used for the adaptive NN and no offline training phase is needed. Experimental results obtained by testing the controller on a research engine indicate an 80% drop of NOx from stoichiometric levels using 10% EGR. Moreover, unburned hydrocarbons drop by 25% due to NN control as compared to the uncontrolled scenario.

Neural network-based output feedback controller for lean operation of spark ignition engines

Spark ignition (SI) engines running at very lean conditions demonstrate significant nonlinear behavior by exhibiting cycle-to-cycle dispersion of heat release even though such operation can significantly reduce NOx emissions and improve fuel efficiency by as much as 5-10%. A suite of neural network (NN) controller without and with reinforcement learning employing output feedback has shown ability to reduce the nonlinear cyclic dispersion observed under lean operating conditions. The neural network controllers consists of three NN: a) A NN observer to estimate the states of the engine such as total fuel and air; b) a second NN for generating virtual input; and c) a third NN for generating actual control input. For reinforcement learning, an additional NN is used as the critic. The uniform ultimate boundedness of all closed-loop signals is demonstrated by using Lyapunov analysis without using the separation principle. Experimental results on a research engine at an equivalence ratio of 0.77 show a drop in NOx emissions by around 98% from stoichiometric levels. A 30% drop in unburned hydrocarbons from uncontrolled case is observed at this equivalence ratio

Turbulent Forced Convection in a Plane Asymmetric Diffuser: Effect of Diffuser Angle

A simulation of two-dimensional turbulent forced convection in a plane asymmetric diffuser with an expansion ratio of 4.7 is performed, and the effect of the diffuser angle on the flow and heat transfer is reported. This geometry is common in many heat exchanging devices, and the turbulent convective heat transfer in it has not been examined. The momentum transport in this geometry, however, has received significant attention already, and the studies show that the results from the v 2 -f type turbulence models provide better agreement with measured velocity distributions than that from the k-s or k-ω turbulence models. In addition, the v 2 -f type turbulence models have been shown to provide good heat transfer results for separated and reattached flows. The k-e-ζ (v 2 -f type) turbulence model is used in this study due to its improved numerical robustness, and the FLUENT-CFD code is used as the simulation platform. User defined functions for the k-e-ζ turbulence model were developed and incorporated into the FLUENT-CFD code, and that process is validated by simulating the flow and the heat transfer in typical benchmark problems and comparing these results with available measurements. This new capability is used to study the effect of the diffuser angle on forced convection in an asymmetric diffuser, and the results show that the angle influences significantly both the flow and the thermal field. The increase in that angle increases the size of the recirculation flow region and enhances the rate of the heat transfer.

Shear-Driven Liquid Film in a Duct

Abstract Two-dimensional flow simulations of shear-driven thin liquid film by turbulent air flow in a duct is performed using the Reynolds Averaged Navier Stokes and continuity equations along with the Volume of Fluid (VOF) model that is part of the FLUENT-CFD code. The purpose of this study is to determine the suitability of using this code/model for predicting reported measurements of shear driven liquid film in a duct. Both a laminar and a turbulent flow models were examined for the liquid film flow region to assess their impact on film thickness and velocity. The Low Reynolds number k-ε turbulence model is utilized for simulating the turbulent air and film flow. Simulated results for the distributions of the air velocity, liquid film velocities along with the liquid film thickness as a function of inlet air and liquid film flow rates are presented. Simulated results show that the film thickness decreases but surface film velocity increases with increasing air flow rate; film thickness and surface film velocity increase with increasing film flow rate; the imposed laminar film flow model produces linear velocity distribution inside the film but the turbulence film flow model produces a nonlinear velocity distribution; the developing thin film influences significantly the air velocity distribution; and these results compare favorably with measured behavior.

Influence of the Combustion Energy Release on Surface Accelerations of an HCCI Engine

Large cyclic variability along with increased combustion noise present in low temperature combustion (LTC) modes of internal combustion engines has driven the need for fast response, robust sensors for diagnostics and feedback control. Accelerometers have been shown as a possible technology for diagnostics and feedback control of advanced LTC operation in internal combustion engines. To make better use of this technology, an improved understanding is necessary of the effect of energy release from the combustion process on engine surface vibrations. This study explores the surface acceleration response for a single-cylinder engine operating with homogeneous charge compression ignition (HCCI) combustion. Preliminary investigation of the engine surface accelerations is conducted using a finite element analysis of the engine cylinder jacket along with consideration of cylindrical modes of the engine cylinder. Measured in-cylinder pressure is utilized as a load input to the FE model to provide an initial comparison of the computed and measured surface accelerations. Additionally, the cylindrical cavity resonant modes of the engine geometry are computed and the in-cylinder pressure frequency content is examined to verify this resonant behavior. Experimental correlations between heat release and surface acceleration metrics are then used to identify specific acceleration frequency bands in which characteristics of morexa0» the combustion heat release process is detected with minimal structural resonant influence. Investigation of a metric capable of indicting combustion phasing is presented. Impact of variations in the combustion energy release process on the surface accelerations is discussed. «xa0less

Online near optimal control of unknown nonaffine systems with application to HCCI engines

Multi-input and multi-output (MIMO) optimal control of unknown nonaffine nonlinear systems is a challenging problem due to the presence of control inputs inside the unknown nonlinearity. In this paper, the optimal control of MIMO nonlinear nonaffine discrete-time systems in input-output form is considered when the internal dynamics are unknown. First, the nonaffine nonlinear system is converted into an affine-like equivalent nonlinear system under the assumption that the higher-order terms are bounded. Next, a forward-in-time Hamilton-Jaccobi-Bellman (HJB) equation-based optimal approach is developed to control the affine-like nonlinear system using neural network (NN). To overcome the need to know the control gain matrix of the affine-like system for the optimal controller, an online identifier is introduced. Lyapunov stability of the overall system including the online identifier shows that the approximate control input approaches the optimal control with a bounded error. Finally, the optimal control approach is applied to the cycle-by-cycle discrete-time representation of the experimentally validated HCCI engine which is represented as a nonaffine nonlinear system. Simulation results are included to demonstrate the efficacy of the approach in presence of actuator disturbances.

Shear Driven Liquid Film in a Duct: Comparison With Measured Results

Two-dimensional flow simulations of shear driven liquid film by turbulent air flow in a duct is performed using the Reynolds Averaged Navier Stokes and continuity equations along with the k-e turbulence model and the Volume of Fluid (VOF) model that are part of the FLUENT-CFD code. The purpose of this effort is to determine the suitability of using this model for predicting the measured results of Wittig et al. [1, 2]. In these studies, optical means are used to measure shear driven liquid film velocities along with the liquid film thickness, and measurements are reported for different air velocities and liquid film flow rates. The simulated results will be compared with measured values, and the suitability of the VOF model for simulating such flow will be assessed in this study.© 2006 ASME

Three-Dimensional Shear Driven Thin Liquid Film in a Duct

Measurements and predictions of three-dimensional shear driven thin liquid films by turbulent air flow in a duct are reported. FLUENT - CFD code is used to perform the numerical simulations and the Reynolds Averaged Navier-Stokes and continuity equations along with the Volume of Fluid (VOF) model and the realizable k-e turbulence model are implemented for this task. Film thickness and width are reported as a function of air flow rate, liquid film volume flow rate and surface tension, and a comparison with preliminary measured results is made. The thickness of the shear driven liquid film is measured using an interferometric technique that makes use of the phase shift between the reflection of incident light from the top and bottom surfaces of the thin liquid film. The spatial resolution is determined based on the spot size of the incident light, which for the current configuration of the transmitting optics is approximately 10 microns. The resulting fringe pattern is imaged using a high-speed imaging camera operating at 2000 frames per second. The technique has proved successful in measuring thickness between 100 and 900 microns in these shear driven films. Simulation results reveal that higher gas flow velocity decreases the film thickness but increases its width, while higher liquid film flow rate increases the film thickness and increases its width. Reasonable comparison appears to exist between preliminary measured and simulated results.© 2006 ASME

Gas and liquid phase transport in pulsed fuel sprays

An infrared extinction technique was used to characterize spatially and temporally the vapor phase of a transient fuel spray. Optical extinction data were taken at two laser wavelengths and compared to obtain fuel vapor partial pressure values averaged over the line-of-sight through the spray. A transient fuel injector spray was characterized at several axial positions with a spatial resolution of 0.125 cm and a temporal resolution of 0.2 ms. With the knowledge of an axisymmetric spray pulse, the line-of-sight averaged results were deconvoluted to obtain spatially resolved vapor pressure data. The same spray was also characterized using phase/Doppler interferometry to obtain droplet size and velocity distributions. Gas phase velocity information, obtained from the behavior of the smallest droplets, was combined with the spatially resolved vapor results to obtain the vapor flux. The results provide a detailed comparison of liquid and vapor transport in a transient spray environment. (Author)