Richard J. Atkinson

Development of a linear alternator-engine for hybrid electric vehicle applications

This paper examines the design and operation of a generation system that utilizes a linear crankless internal combustion engine in conjunction with a linear alternator. This system directly utilizes the linear motion of the piston to drive the alternator rather than first converting to rotary motion. The result is a more compact, reliable, and efficient unit as the system has only one moving part, making the system ideal for use in series hybrid electric vehicles. This paper describes the overall system design as well as the subsystems including the engine and alternator. A dynamic simulation is then presented which utilizes the model developed to determine the output characteristics of the system. The prototype system was successfully tested, and experimental results are also included.

Numerical Simulation of a Two-Stroke Linear Engine-Alternator Combination

Series hybrid electric vehicles (HEVs) require powerplants that can generate electrical energy without specifically requiring rotary input shaft motion. A small-bore working prototype of a two-stroke spark ignited linear engine-alternator combination has been designed, constructed and tested and has been found to produce as much as 316W of electrical energy. This engine consists of two opposed pistons (of 36 mm diameter) linked by a connecting rod with a permanent magnet alternator arranged on the reciprocating shaft. This paper presents the numerical modeling of the operation of the linear engine. The piston motion of the linear engine is not mechanically defined: it rather results from the balance of the in-cylinder pressures, inertia, friction, and the load applied to the shaft by the alternator, along with history effects from the previous cycle. The engine computational model combines dynamic and thermodynamic analyses. The dynamic analysis performed consists of an evaluation of the frictional forces and the load (in this case the alternator load) across the full operating cycle of the engine. The thermodynamic analysis consists of an evaluation of each process that characterizes the engine cycle, including scavenging, compression, combustion and expansion, based on the first law of thermodynamics. Since the modeled engine was crankshaftless, a time-based Wiebe function (as opposed to a conventional crank angle-based approach) was used to express the mass fraction burned for the combustion process, while the combustion model used was a single-zone model. To render the model useful, the parameters used were based on experimental data obtained from the working example, including instantaneous shaft position, velocity and in-cylinder pressure. Also, a parametric study was performed to predict the behavior of the engine over a wide operating range, given variations in fuel combustion properties, the reciprocating mass of the piston shaft assembly, frictional load and the externally applied electrical load. INTRODUCTION AND LITERATURE REVIEW Free-piston engines have been a subject of research and development for several decades. A recent review of freepiston engine concepts has been conducted by Achten [16]. Free-piston engines utilizing internal combustion (as opposed to the external combustion Stirling engine, which suffers from poor power density) have their origin in the 1920’s when R. Pescara [1] patented their use as air compressors. Junkers in Germany developed a freepiston engine for use in German submarines in World War II. The French SIGMA free-piston gasifier saw service for decades in stationary power generation. The use of free-piston engines in automotive application was most heavily promoted in the period 1952 to 1961, when both General Motors and Ford Motor Company produced running prototypes [2], [3]. In both cases these engines were two-stroke, opposed piston spark ignited engines with combustion bounce/compression chambers. These engines where used as gasifiers to generate hot gases to drive exhaust turbines through which energy would be extracted. Development efforts largely ceased by the 1960’s as turbine powered vehicles were increasingly viewed as not commercially viable. The elimination of the crankshaft mechanism in free-piston engines provides potential for reduction in mechanical losses. Due to the fact that the piston is not constrained in a free-piston engine, the piston motion is not prescribed, and it varies from one operating regime to another. Characteristic of the free-piston engines is the fact that they do not have a flywheel. As a result, they do not accumulate energy from the previous cycles for the subsequent cycles, except in terms of achieving a greater or lesser stroke associated with greater or lesser gas compression energy. A linear engine operates somewhat similarly to a free-piston engine, the only difference being that the reciprocating assembly consists of two pistons connected by a common connecting rod, with each piston operating in its own cylinder.

Testing of a Heavy Heavy-Duty Diesel Engine Schedule for Representative Measurement of Emissions

Abstract The Advanced Collaborative Emissions Study (ACES) program required the use of representative heavy-duty diesel engine activity. This need resulted in an engine test schedule creation program, and a schedule of engine modes representative of modern truck usage was developed based on data collected from engines in trucks operated through the heavy heavy-duty diesel truck (HHDDT) chassis schedule. The ACES test schedule included four active modes of truck operation including creep, transient, cruise, and high-speed cruise (HHDDT_S). This paper focuses on Phase 2 of the program, which was to validate and demonstrate the use of the ACES modes in a test cell. Preliminary testing was performed using a 1992 Detroit Diesel Corporation heavy heavy-duty diesel engine (HHDDE) on only the transient mode. On the basis of these results, each mode was modified slightly to suit implementation in a test cell. The locations of “closed throttle” points in the modes were determined through careful examination of the data. These closed throttle points were simulated during testing by adding negative set point torque values to the input file. After modification, all modes were tested during a final ACES modes demonstration period using a 2004 Cummins ISM HHDDE, obtaining three runs for each mode. During testing, carbon monoxide (CO), carbon dioxide (CO2), nitrogen oxides (NOx), particulate matter (PM), and hydrocarbon (HC) emissions were measured, and engine control unit (ECU) data were recorded. The new ACES modes did not adopt the Federal Test Procedure (FTP) regression criteria. New regression criteria for acceptability of a run were determined for each mode using the data obtained during testing.

A CFR1065-Compliant Transportable/On-Road Low Emissions Measurement Laboratory With Dual Primary Full-Flow Dilution Tunnels

In 2007, certification standards for heavy duty diesel particulate matter (PM) emissions were reduced from 0.1g/bhp-hr to 0.01g/bhp-hr, representing an order of magnitude reduction in pollutant level. Coincident with these standards revisions are refinements to test procedures that target reductions in measurement uncertainties. The 2007 U.S. Environmental Protection Agency (US EPA) specifications, as defined in 40 CFR parts 86, and US EPA 2010 specifications, as defined in CFR 1065, require significant updates to established laboratory measurement systems and test procedures. Moreover, additional regulatory standards pertaining to in-use compliance of heavy duty diesel engines will significantly impact the future of heavy duty diesel emissions measurement. As a result of the reduced emission production levels, demand for ‘real-world’ emissions measurements, and subsequent development and evaluation of on-board emissions measurement systems, West Virginia University’s Center for Alternative Fuels, Engines, and Emissions (CAFEE) has designed and constructed, with support from the U.S. Department of Energy (DOE), the ‘next level’ transportable dual primary full-flow dilution tunnel emissions measurement laboratory. The objective of this project was to build a mobile emissions measurement laboratory, of engine test cell quality, that is capable of measuring regulated and non-regulated emissions, and meets US EPA 2007 and 2010 specifications. A thirty-foot long cargo container was constructed to house a portable emissions measurement facility, comprised of a High Efficiency Particulate Air (HEPA) primary dilution unit, two primary full-flow dilution tunnels, a subsonic venturi, a secondary particulate matter sampling system, a gaseous emissions analytical bench instrumentation system, a computer based data acquisition (DAQ) and control system, full air conditioning and ventilation system, and chassis dynamometer control systems. Dual tunnels, of 18 inches ID and 20 feet long provide dedicated measurement capability for both lower PM vehicles, as well as legacy diesel fueled vehicles. This provision reduces tunnel history effects between test programs which address differing exhaust composition and PM loading. The laboratory grade analytical system can be transported to virtually any location with a demand for emissions testing, either with or without WVU’s transportable medium or heavy duty chassis dynamometers. Alternatively, the system can be loaded onto a flatbed trailer in order to test emissions while a vehicle is operated over the road. This paper describes each sub-system of this transportable laboratory in the aspect of specifications and design considerations, and presents results of qualification tests on the laboratory.Copyright

Turbocharging a bi-fuel engine for performance equivalent to gasoline

A bi-fuel engine capable of operating either on compressed natural gas (CNG) or gasoline is being developed for the transition to alternate fuel usage. A Saturn 1.9 liter 4-cylinder engine was selected as a base powerplant. A turbocharger was installed to increase the density of the intake charge and thereby regain the volumetric efficiency lost with CNG. Reductions from baseline in hydrocarbon and carbon dioxide emissions were achieved at power levels equivalent to and slightly higher than the baseline. Brake thermal efficiency values were not significantly different in any case. 8 refs., 12 figs.

A Controller for a Spark Ignition Engine with Bi-Fuel Capability

A bi-fuel engine with the ability to run optimally on both compressed natural gas (CNG) and gasoline is being developed. The engine control system described here employs adaptive closed-loop control optimizing fuel delivery and spark timing for both fuels. Ic-cylinder pressure was measured in the experimental engine using piezoelectric pressure transducers. Reductions in engine-out emissions, as compared with stock operations, were observed when using the controller with gasoline. Further reductions in emissions were achieved with CNG operation, due to properties of the fuel. An improvement in engine stability was also realized with the controller. 14 refs., 12 figs.

Development of an Interface Method for Implementing Road Grade in Chassis Dynamometer Testing

Mobile source emissions inventory data from heavy-duty on-road vehicles are traditionally obtained using three methods: engine dynamometer, chassis dynamometer, or in-use vehicle driving. Engine dynamometer testing provides for the greatest control and highest accuracy but requires the most time and can be cost prohibitive when obtaining emissions from many in-use engines. In-use emissions collection is a relatively inexpensive and rapid method of obtaining real-world data, but this method is relatively new and is not regulated by any Federal or international regulations as of yet and accuracy of the data from these devices has not been established. Chassis dynamometer-based testing provides for the means of obtaining a large sample of data from in-use vehicles in a controlled environment. However, existing chassis dynamometer cycles assume a level road surface with no grade. In a chassis dynamometer test cycle, a simple line trace is used to represent the desired vehicle speed on a video monitor for the driver to follow. A second line trace is overlaid on the first to indicate the actual vehicle speed on the dynamometer and the drive adjusts the vehicle speed to match the scheduled speed as closely as he is able. Experienced chassis dynamometer test drivers are able to look at the desired speed and anticipate the required gearshifts during the testing. However, to account for road grade in a chassis dynamometer test schedule, the driver of the vehicle will require additional cues. Also, drivers may not drive a vehicle while following a trace in the same way that they drive on the road. To implement grade and inject a sense of the real world in a chassis dynamometer test cycle, a virtual reality interface has been developed to employ images of a roadway with feedback between the driver’s performance and the image. As a first step to implementing grade, a level road surface using a virtual reality interface was emulated using an in-house developed software package to present images of roadways, including traffic control signals and constraints due to traffic congestion. In the virtual reality execution, the driver perceives the position of the vehicle relative to traffic signals and other traffic cues. An initial investigation into the effect of road grade using the conventional line trace method is shown and then use of the virtual reality approach is compared with the conventional line trace. The results from the study shows that an experienced driver can use the virtual reality interface with similar emission results as the conventional line trace method.Copyright

Novel NOx Emission Reduction Technology for Diesel Marine Engines

Emissions from diesel marine engines are significant contributors to the emissions inventories of commercial ports. Prior to 1998, these emissions were unregulated and current EPA regulations apply predominantly to new engines. Considering that the useful life of marine engines in work vessels, such as tugboats, may be 20 years or longer, retrofit emission reduction technologies are needed for these legacy engines. Oxides of nitrogen (NOx ) have negative health and environmental impacts and are difficult to reduce substantially without aftertreatment. A scrubber system for NOx reduction was proposed; the presented research focuses on the verification of operating principles and the quantification of possible NOx reduction from this system. Major elements of the proposed scrubber system are exhaust heat exchangers, a catalyzed particulate filter (CPF), a diesel oxidation catalyst (DOC), and a packed bed wet scrubber. The system works on the principle of absorption of NOx species into water. The majority of engine-out NOx is in the form of nitric oxide (NO) which is relatively insoluble in water. A CPF and a DOC are utilized to convert up to 80% of the NO into nitrogen dioxide, NO2 . NO2 and NO exist in equilibrium with N2 O3 and N2 O4 , species of NOx that are highly soluble in water. The use of a CPF and DOC also reduces carbon monoxide, hydrocarbons, and particulate matter, reducing possible scrubber contamination. The scrubber liquor operates on a closed loop with zero discharge, its final composition is weak nitric acid; a byproduct of capturing the NOx . Research to support this design was conducted on a Mack E7 298 kW, 12 liter engine operating over 8 steady state points. Modal NOx absorption ranged from 4–66%. Cycle average NOx absorption ranged from 15–58%. It was concluded that NOx absorption varies with gas residence time, absorption surface area, temperature, and NOx concentration. Separately, a system was constructed and operated to convert the stored concentrated NOx into diatomic nitrogen, carbon dioxide, and water.Copyright

System Model for Selective NOx Recirculation (SNR) to be Used in Stationary Lean-Burn Natural Gas Engines

A model for a possible system to implement Selective NOx Recirculation (SNR) technology for stationary lean-burn natural gas engines was developed. SNR is a NOx (NOx includes the various oxides of nitrogen found in an exhaust stream) removal after-treatment technology with four phases; cooling the hot exhaust gas, NOx adsorption onto a sorbent material, periodic NOx desorption using heat, and NOx decomposition within the combustion process. This paper presents the model, summarizes the research used to develop the model, and presents model output. NOx decomposition in the combustion process was investigated by injecting nitric oxide (NO) into the intake of a Cummins L10G natural gas fueled spark-ignited engine (210 kW at 2100 rpm). Experimental campaigns were conducted during lean-burn and rich-burn operation to quantify in-cylinder NOx decomposition. Data previously published suggest that rich burn is essential for adequate NOx decomposition and that lean burn was ineffective. The NOx adsorption/desorption characteristics of the sorbent material were quantified using a bench top adsorption system equipped with four thermocouples, an in-line heater, a mass flow controller and a Rosemont Analytical NOx analyzer. The sorbent chamber was filled with activated carbon sorbent material. Extensive testing of the adsorption characteristics using 500 ppm NO (balance nitrogen) from a pressure tank yielded a mass percent of .0005 NO to carbon. These results suggested that unacceptably large adsorbers would be needed in industrial applications. However, further measurement using real exhaust showed a loading of 0.65 mass percent of NO to carbon. The presence of oxygen and water are implicated in this improved adsorption. This scaled system considered the heating rates (for desorption) and cooling rates (for adsorption) for the bed at the time when desorption and adsorption processes were initiated. An adsorption/desorption model that considered gas temperature and heat and mass transfer was formulated based on these data. A simplified linear driving force model was developed to predict NOx adsorption into the sorbent material as cooled exhaust passed over fresh sorbent material.Copyright

Nitric Oxide Conversion in a Spark Ignited Natural Gas Engine

Nitric Oxide Conversion in a Spark Ignited Natural Gas Engine Matthew M. Swartz Reducing NOX emissions from natural gas engines has become increasingly important from an environmental standpoint. A large percentage of stationary engine applications are natural gas fueled. The cleanest of these large bore engines currently produce on the order of one gram of NOX per brake-horsepower hour (g/bhp-hr) of work done. The goal of this work is to reduce these emissions to 0.1 g/bhp-hr levels. Selective NOX Recirculation (SNR), a technology which will help achieve these 0.1 g/bhp-hr levels, is currently being studied at West Virginia University. SNR has been proven in gasoline and diesel engines, with up to 90% NOX conversion rates being achieved, but not much is known about its overall efficiencies when used with natural gas engines. This technique involves adsorbing NOX from an exhaust stream, then selectively desorbing the NOX into a concentrated NOX stream, which is fed back into the engine’s intake, thereby converting a percentage of the concentrated NOX stream into harmless gases. Understanding the NO (a component of NOX) conversion process plays a major role in optimizing the SNR technology. The emphasis of this thesis is on the unique chemical kinetic modeling problem that occurs with high concentrations of NOX in the intake air of a spark ignited natural gas engine with SNR. NO conversion experiments were performed on a Cummins natural gas engine, and CHEMKIN, a chemical kinetic solver software package was used to simulate this process. A closed homogeneous batch reactor model was used to model the concentration of NOX versus time for an initial mixture of NO, O2, N2, and CH4. A zerodimensional model was used to model the NOX conversion properties of a natural gas engine. The molar fraction of NOX was monitored on small time scales, on the order of the time for complete combustion, and on large time scales, on the order of minutes. Predicting the mole fraction of NOX as a function of time using a closed homogenous batch reactor model on small time scales allowed estimates of the percentage of NO converted locally during the combustion process. These percentages were then compared with experimental values, which were acquired from a Cummins 10 liter spark ignited natural gas engine. The model predicted conversion rates, based on percentage mass conversion, varying between 20% and 26%, and experiments showed conversion rates between 18% and 23% for a constant intake NO concentration of 25,000 ppm. Predicting NOX concentration on large time scales, steady state, confirmed that the NOX conversion phenomenon, over the in-cylinder time period in which combustion temperatures are maintained, is rate limited rather than equilibrium limited. This gives insight into how to maximize these conversion efficiencies for the SNR process. If the process is rate limited rather than equilibrium limited combustion temperature will play a larger role in NO conversion. Once the experimental setup was modeled, CHEMKIN was then used to predict the effects of EGR and varying air-to-fuel (A/F) ratios on SNR efficiencies. CHEMKIN demonstrated that under ideal conditions, increasing the air to fuel ratio and adding EGR will increase NO conversion efficiency by up to 90%.