Predictive modeling and testing of a diesel derived solid oxide fuel cell tail gas spark-ignition engine
Date
2020
Authors
Countie, Matthew, author
Olsen, Daniel, advisor
Windom, Brett, committee member
Baker, Daniel, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Solid oxide fuel cell systems are being developed with total system efficiency targets over 70%. One approach is to provide excess fuel to the solid oxide fuel cell and develop an engine to provide power for mechanical and electrical equipment using exhaust gas from the fuel cell anode (tailgas). This tailgas contains hydrogen, carbon monoxide, methane, water, and carbon dioxide. Compared to natural gas the tailgas fuel has suppressed flame speeds, an extremely small lower heating value, and a low air-fuel ratio due to the presence of large amounts of oxidation products. A predictive model created in GT-Power was used to design an engine that can produce 14kW on tailgas fuel with a brake efficiency η>30%. The model base is an existing Kohler diesel engine. The diesel engine was modeled in GT-Power and validated to within 1% at the anticipated operating point. Using custom combustion models developed from testing several different tailgas blends in a CFR engine, several different engine conversions were modeled to explore different pathways to 30% brake efficiency. Design variations include Miller cycles, turbocharging, compression ratio, and fuel pre-treatment to increase reactivity. Once design parameters were established, an operation envelope was created to identify knock limits and maximum brake efficiency timing. These models helped guide the development of a physical prototype engine that was built and installed at the CSU Powerhouse Energy Campus. The prototype engine ran with simulated anode tailgas up to a maximum power level of 7.42 kW and a maximum brake efficiency of 27.34%, achieving 53% of the load target, and 91% of the efficiency target. The timings identified by GT-Power to be the point of maximum brake efficiency and knock initiation were tested at four different speeds on the prototype engine. After data collection, using the experimental power, engine speed, and ignition timing as initial conditions, the model is rerun. The accuracy of the models' prediction capability is tested by using these initial conditions to generate additional model output to compare with measured data. At low speeds and advanced ignition timings, the model matched well, within 10% on almost all metrics, but at retarded timings and high engine speeds, the model began to deviate in most parameters, especially overpredicting exhaust temperature and pressure. The discrepancies between model results and experimental data are discussed in detail. Model and experimental data matched well at advanced timings and low speeds, but deviated significantly at retarded timings and high speeds.
Description
Rights Access
Subject
engines
internal combustion engine
dilute fuels
solid oxide fuel cell
GT-power