Search Results

A procedure has been developed to build system level predictive models that incorporate physical laws as well as information derived from experimental data. In particular a soot model was developed, trained and tested using experimental data. It was seen that the model could fit available experimental data given sufficient training time. Future accuracy on data points not encountered during training was estimated and seen to be good. The approach relies on the physical phenomena predicted by an existing system level phenomenological soot model coupled with ‘weights’ which use experimental data to adjust the predicted physical sub-model parameters to fit the data. This approach has developed from attempts at incorporating physical phenomena into neural networks for predicting emissions. Model training uses neural network training concepts.

Ignition dwell is defined as the interval between end of fuel injection and start of combustion in early injection diesel combustion that exhibits HCCI-like characteristics. In this project, the impact of in-cylinder temperature and fuel-air mixing on the ignition dwell was investigated. The engine cycle was simulated using the 3-D CFD code KIVA-3V. Work done by Klingbeil (2002) has shown that ignition dwell allows more time for fuel and air to mix and drastically reduces emissions of NOX and particulate matter. Temperature is known to have a direct impact on the duration of ignition dwell. However, initial fuel-air distribution and mixing (i.e. at the end of fuel injection) may also impact the duration of ignition dwell. To investigate this, variations in EGR, fuel injection timing, engine valve actuation and swirl were simulated. The aim was to use these techniques to generate varying levels of fuel-air mixing and to check if ignition dwell was affected.

Low-temperature combustion concepts that utilize cooled EGR, early/retarded injection, high swirl ratios, and modest compression ratios have recently received considerable attention. To understand the combustion and, in particular, the soot formation process under these operating conditions, a modeling study was carried out using the KIVA-3V code with an improved phenomenological soot model. This multi-step soot model includes particle inception, surface growth, surface oxidation, and particle coagulation. Additional models include a piston-ring crevice model, the KH/RT spray breakup model, a droplet wall impingement model, a wall heat transfer model, and the RNG k-ε turbulence model. The Shell model was used to simulate the ignition process, and a laminar-and-turbulent characteristic time combustion model was used for the post-ignition combustion process.

The strategy of variable Intake Valve Closure (IVC) timing, as a means to improve performance and emission characteristics, has gained much acceptance in gasoline engines; yet, it has not been explored extensively in diesel engines. In this study, genetic algorithms are used in conjunction with the multi-dimensional engine simulation code KIVA-3V to investigate the optimum operating variables for a typical heavy-duty diesel engine working with late IVC. The effects of start-of-injection timing, injection duration and exhaust gas recirculation were investigated along with the intake valve closure timing. The results show that appreciable reductions in NOx+HC (∼82%), soot (∼48%) and BSFC (∼7.4%) are possible through this strategy, as compared to a baseline diesel case of (NOx+HC) = 9.48g/kW-hr, soot = 0.17 g/kW-hr and BSFC = 204 g-f/kW-hr. The additional consideration of double injections helps to reduce the high rates of pressure rise observed in a single injection scheme.

In this paper, optimized single and double injection schemes were found using multi-dimensional engine simulation software (KIVA-3V) and a micro-genetic algorithm for a heavy duty diesel engine. The engine operating condition considered was at 1737 rev/min and 57 % load. The engine simulation code was validated using an engine equipped with a hydraulic-electronically controlled unit injector (HEUI) system. Five important parameters were used for the optimization - boost pressure, EGR rate, start-of-injection timing, fraction of fuel in the first pulse and dwell angle between first and second pulses. The optimum results for the single injection scheme showed significant improvements for the soot and NOx emissions. The start of injection timing was found to be very early, which suggests HCCI-like combustion. Optimized soot and NOx emissions were reduced to 0.005 g/kW-hr and 1.33 g/kW-hr, respectively, for the single injection scheme.

The effect of injection rate shape on spray evolution and emission characteristics is investigated and a methodology for active control of fuel injection is proposed. Extensive validation of advanced vaporization and primary jet breakup models was performed with experimental data before studying the effects of systematic changes of injection rate shape. Excellent agreement with the experiments was obtained for liquid and vapor penetration lengths, over a broad range of gas densities and temperatures. Also the predicted flame lift-off lengths of reacting diesel fuel sprays were in good agreement with the experiments. After the validation of the models, well-defined rate shapes were used to study the effect of injection rate shape on liquid and vapor penetration, flame lift-off lengths and emission characteristics.

Neural networks are useful tools for optimization studies since they are very fast, so that while capturing the accuracy of multi-dimensional CFD calculations or experimental data, they can be run numerous times as required by many optimization techniques. This paper describes how a set of neural networks trained on a multi-dimensional CFD code to predict pressure, temperature, heat flux, torque and emissions, have been used by a genetic algorithm in combination with a hill-climbing type algorithm to optimize operating parameters of a diesel engine over the entire speed-torque map of the engine. The optimized parameters are mass of fuel injected per cycle, shape of the injection profile for dual split injection, start of injection, EGR level and boost pressure. These have been optimized for minimum emissions. Another set of neural networks have been trained to predict the optimized parameters, based on the speed-torque point of the engine.

Neural networks have been used for engine computations in the recent past. One reason for using neural networks is to capture the accuracy of multi-dimensional CFD calculations or experimental data while saving computational time, so that system simulations can be performed within a reasonable time frame. This paper describes three methods to improve upon neural network predictions. Improvement is demonstrated for in-cylinder pressure predictions in particular. The first method incorporates a physical combustion model within the transfer function of the neural network, so that the network predictions incorporate physical relationships as well as mathematical models to fit the data. The second method shows how partitioning the data into different regimes based on different physical processes, and training different networks for different regimes, improves the accuracy of predictions.

A cylinder model was developed using artificial neural networks (ANN). The cylinder model utilized the trained ANN models to predict engine parameters including cylinder pressures, cylinder temperatures, cylinder wall heat transfer, NOx and soot emissions. The ANN models were trained to approximate CFD simulation results of an engine. The ANN cylinder model was then applied to predict engine performance and emissions over the standard heavy-duty FTP transient cycle. The engine responses varying over the engine speed and torque range were simulated in the course of the transient test cycle. It was demonstrated that the ANN cylinder model is capable of simulating the characteristics of the engine operating under transient conditions reasonably well.

Optimizations were performed on a Heavy-Duty diesel engine equipped with a conventional electronic unit injector in order to minimize fuel consumption, and emissions of NOx and particulate matter. A low speed light load case and a high speed light load case were optimized with these considerations in mind. Exhaustive parametric studies were performed in order to find sets of operating conditions that resulted in low emissions and high fuel economy. It was found for the low speed light load case (Mode 2, 25% load and 821 rev/min) that low emissions operating conditions existed at either very early or very late start-of-injection timings and high EGR (PM = 0.018 g/kW-hr, NOx + HC = 1.493 g/kW-hr with SOI = -21 degrees ATDC, 48% EGR; or 0.085 g/kW-hr PM, 1.02 g/kW-hr NOx with SOI = 4 degrees ATDC, 39% EGR).

To overcome the trade-off between NOx and particulate emissions for future diesel vehicles and engines it is necessary to seek methods to lower pollutant emissions. The desired simultaneous improvement in fuel efficiency for future DI (Direct Injection) diesels is also a difficult challenge due to the combustion modifications that will be required to meet the exhaust emission mandates. This study demonstrates the emission reduction capability of split injections, EGR (Exhaust Gas Recirculation), and other parameters on a High Speed Direct Injection (HSDI) diesel engine equipped with a common rail injection system using an RSM (Response Surface Method) optimization method. The optimizations were conducted at 1757 rev/min, 45% load. Six factors were considered for the optimization, namely the EGR rate, SOI (Start of Injection), intake boost pressure, and injection pressure, the percentage of fuel in the first injection, and the dwell between injections.

Current automotive diesel engine research is motivated by the need to meet more-and-more strict emission regulations. The major target for future HSDI combustion research and development is to find the most effective ways of reducing the soot particulate and NOx emissions to the levels required by future emission regulations. Recently, a variety of statistical optimization tools have been proposed to optimize engine-operating conditions for emissions reduction. In this study, a micro-genetic algorithm technique, which locates a global optimum via the law of “the survival of the fittest”, was applied to a high-speed, direct-injection, single-cylinder (HSDI) diesel engine. The engine operating condition considered single-injection operation using a common-rail fuel injection system was at 1757 rev/min and 45% load.

Artificial neural networks (ANN) have been recognized as universal approximators for nonlinear continuous functions and actively applied in engine research in recent years [1, 2, 3, 4, 5, 6, 7 and 8]. This paper describes the methodology and results of using the ANN to model a turbocharged DI diesel engine. The engine was simulated using the CFD code (KIVA-ERC) over a wide range of operating conditions, and numerical simulation results were used to train the ANN. An efficient data collection methodology using the Design of Experiments (DOE) techniques was developed to select the most characteristic engine operating conditions and hence the most informative data to train the ANN. This approach minimizes the time and cost of collecting training data from either computational or experimental resources. The trained ANN was then used to predict engine parameters such as cylinder pressure, cylinder temperature, NOx and soot emissions, and cylinder heat transfer.

A micro-genetic algorithm (μGA) optimization method was applied to a heavy-duty, direct-injected diesel engine via an automated test bed system. The goal of this application was to demonstrate the feasibility and advantages of automated optimization experiments. With the genetic algorithm, no user input was required other than the factors of interest and their allowable ranges. This means that once the routine was initiated, it was essentially run undisturbed until a preset objective level was reached or a preset number of generations had been run. The automated μGA was successfully demonstrated at all points of the six-mode transient cycle simulation, excluding idle. To accomplish the automated experiments, an automated testing system was developed around a Caterpillar single-cylinder diesel engine.

A computational optimization study was performed for a direct-injection diesel engine using a recently developed 1-D-KIVA3v-GA (1-Dimensional-KIVA3v-Genetic Algorithm) computer code. The code performs a full engine cycle simulation within the framework of a genetic algorithm (GA) code. Design fitness is determined using a 1-D (one-dimensional) gas dynamics code for the simulation of the gas exchange process, coupled with the KIVA3v code for three-dimensional simulations of spray, combustion and emissions formation. The 1-D-KIVA3v-GA methodology was used to simultaneously investigate the effect of eight engine input parameters on emissions and performance for four cases, which include cases at 2500 RPM and 1000 RPM, with both simulated at high-load and low-load conditions.

The recently developed KIVA-GA computer code was used in the current study to optimize the combustion chamber geometry of a heavy -duty diesel truck engine and a high-speed direct-injection (HSDI) small-bore diesel engine. KIVA-GA performs engine simulations within the framework of a genetic algorithm (GA) global optimization code. Design fitness was determined using a modified version of the KIVA-3V code, which calculates the spray, combustion, and emissions formation processes. The measure of design fitness includes NOx, unburned HC, and soot emissions, as well as fuel consumption. The simultaneous minimization of these factors was the ultimate goal. The KIVA-GA methodology was used to optimize the engine performance using nine input variables simultaneously. Three chamber geometry related variables were used along with six other variables, which were thought to have significant interaction with the chamber geometry.

Experimental data is used in conjunction with multi-dimensional modeling in a modified version of the KIVA-3V code to characterize the emissions behavior of a high-speed, direct-injection diesel engine. Injection pressure and EGR are varied across a range of typical small-bore diesel operating conditions and the resulting soot-NOx tradeoff is analyzed. Good agreement is obtained between experimental and modeling trends; the HSDI engine shows increasing soot and decreasing NOx with higher EGR and lower injection pressure. The model also indicates that most of the NOx is formed in the region where the bulk of the initial heat release first takes place, both for zero and high EGR cases. The mechanism of NOx reduction with high EGR is shown to be primarily through a decrease in thermal NOx formation rate.

A study of the combined use of split injections, EGR, and flexible boosting was conducted. Statistical optimization of the engine operating parameters was accomplished using a new response surface method. The objective of the study was to demonstrate the emissions and fuel consumption capabilities of a state-of-the-art heavy -duty diesel engine when using split injections, EGR, and flexible boosting over a wide range of engine operating conditions. Previous studies have indicated that multiple injections with EGR can provide substantial simultaneous reductions in emissions of particulate and NOx from heavy-duty diesel engines, but careful optimization of the operating parameters is necessary in order to receive the full benefit of these combustion control techniques. Similarly, boost has been shown to be an important parameter to optimize. During the experiments, an instrumented single-cylinder heavy -duty diesel engine was used.

A computational optimization study is performed for a heavy-duty direct-injection diesel engine using the recently developed KIVA-GA computer code. KIVA-GA performs full cycle engine simulations within the framework of a Genetic Algorithm (GA) global optimization code. Design fitness is determined using a one-dimensional gas -dynamics code for calculation of the gas exchange process, and a three-dimensional CFD code based on KIVA-3V for spray, combustion and emissions formation. The performance of the present Genetic Algorithm is demonstrated using a test problem with a multi-modal analytic function in which the optimum is known a priori. The KIVA-GA methodology is next used to simultaneously investigate the effects of six engine input parameters on emissions and performance for a high speed, medium load operating point for which baseline experimental validation data is available.

A study of statistical optimization of engine operating parameters was conducted. The objective of the study was to develop a strategy to efficiently optimize operating parameters of diesel engines with multiple injection and EGR capabilities. Previous studies have indicated that multiple injections with EGR can provide substantial simultaneous reductions in emissions of particulate and NOx from heavy-duty diesel engines, but careful optimization of the operating parameters is necessary in order to receive the full benefit of these combustion control techniques. The goal of the present study was to optimize the control parameters to reduce emissions and brake specific fuel consumption. An instrumented single-cylinder heavy-duty diesel engine was used with a prototype mechanically actuated (cam driven) fuel injection system.