<div>Plug-in hybrid electric vehicles (PHEVs) are being considered as an alternative to conventional medium-duty (MD) and heavy-duty (HD) commercial vehicles to reduce fuel consumption and tailpipe emissions. Lithium ion batteries, which are used in PHEVs due to their high energy density, are expensive. The battery contributes significantly towards the life-cycle cost of MD/HD PHEVs, as these vehicles, due to high mass and aggressive battery usage, require multiple battery replacements over their lifetime. Smaller batteries increase the fuel consumption and need more replacements, while bigger batteries increase the initial system cost. Powertrain design from a life-cycle cost perspective is required to explore this trade-off and maximize the economic gains obtained from PHEVs. </div><div><br></div><div>Powertrain design entails component sizing, control strategy selection as well as architecture selection. Different powertrain designs yield different lifetime economic gains. A variety of applications exist for MD/HD vehicles, which differ in their ways of powertrain usage, due to variations in required acceleration, available braking, and average and maximum speeds. Therefore, different powertrain designs are needed depending on the application and usage scenario. The powertrain design space needs to be explored, and solutions that maximize the economic gains within the specified constraints need to be chosen.</div><div><br></div><div>This dissertation compares the economic viability of two PHEV applications (MD Truck and HD Transit Bus), with options of series and parallel hybrid architectures, over multiple drivecycles, for four economic scenarios (years 2015, 2020, 2025 and 2030). It is shown that hybridizing the transit bus achieves payback sooner than hybridizing the truck. Further, the results for the transit bus application, over the Manhattan drivecycle, show that implementation of the parallel architecture is economically viable in the 2015(present) scenario, while the series architecture becomes viable in 2020, due to significantly lower initial costs involved in the parallel architecture.</div><div><br></div><div>A methodology to select a solution out of the explored design space that maximizes the economic gains is demonstrated. Variations in the economic and vehicle usage conditions for which this solution is designed, can be expected. It is therefore necessary to check the robustness of this solution to change in external factors such as vehicle mass, annual vehicle miles travelled (AVMT), component and fuel costs. It is shown that the economic gains are affected by the battery cost, fuel cost, AVMT and vehicle mass, while the number of battery replacements are affected by AVMT and vehicle mass. </div><div><br></div><div>A probability-based approach is demonstrated to obtain confidence in the economic and battery life predictions. Specifically, probability-based variations are provided to variables such as miles traveled between recharge, recharge C-rate and battery temperature. It is shown that battery life is affected the most by battery temperature.</div><div><br></div><div>A battery heating/cooling system is required to maintain constant battery temperature of operation during all seasons, but these systems incur additional fuel costs. A framework that utilizes just the Coefficient of Performance (COP) of the heating/cooling system to calculate the excess fuel cost is proposed and demonstrated. An increase of 0.9-1.8\% in fuel consumption is shown, depending on the drivecycle and ambient temperature.</div><div><br></div><div>Further, the well-to-wheel (WTW) fuel-cycle emissions from conventional and PHEV transit buses operating in Indiana and California are assessed using the ``Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation'' (GREET) Model 2017, developed by Argonne National Labs. It is shown that 59% and 63% greenhouse gas (GHG) reductions can be achieved in Indiana and California respectively, along with reduction in carbon monoxide (CO), nitrogen oxides NOx, particulate matter with diameter less than 2.5 microns PM2.5 and volatile organic compounds (VOC) emissions for both the states. However, an increase in sulfur oxides SOx emissions for both the states, and particulate matter with diameter less than 10 microns PM10 increase for Indiana, are observed. </div><div><br></div>

<p> One primary focus on internal combustion engines is that these engines create multiple harmful exhaust gases that can cause damage to the environment. There are a number of advanced strategies that are currently being investigated to help reduce the amount of these harmful emissions that are emitted from IC engines. One such method of reducing harmful emission gases focuses on the three-way-catalyst. A three-way-catalyst (TWC) is an exhaust emission control device that is designed in such a way to take harmful exhaust gases and convert them into less harmful gases through various chemical reactions within the TWC. To help further the reduction of these harmful gases in the TWC, a novel two-loop control and estimation strategy is used. This control and estimation strategy involves the use of two loops with an inner-loop controller, outer-loop robust controller, and an estimator in the outer-loop. The estimator consists of a TWC model and an extended Kalman filter which is used to estimate the fractional oxidation state (FOS) of the TWC. This estimated FOS is then used by the robust controller, along with other parameters, to produce a desired engine lambda reference signal, λup. This desired lambda signal is then used by the inner-loop controller to control the engine lambda. Accurate control of lambda is important because the air-fuel-ratio range for a TWC to effectively achieve oxidation and reduction simultaneously is extremely narrow. Another primary focus in the field of internal combustion engines is designing and tuning advanced models within GT-Power that can accurately predict what will happen when running an actual engine. Designing, troubleshooting, and testing a GT-Power model is an extensive but rewarding process. Creating an accurate engine model can not only provide one with primary engine data that is also measurable in a test cell, but can also provide insight into some of the intricate processes and nature of the engine that are difficult or impossible to physically measure. Cummins has an extensive process of tuning GT-Power engine models. This process include items such as initial model calibrations, model discretizations, turbocharger tunings, and other items. Some of these processes are used to calibrate both Cummins Power Systems Business Unit engines as well as a Purdue B6.7N natural gas engine. </p>