Master of Science

thesisMetabolic gas analysis systems are important devices that are used to analyze respiratory gas exchange including volumetric flow rates and oxygen and carbon dioxide concentrations. This information provides useful insights into metabolic function. Traditionally, these systems were limited by their size and the functional requirements of the gas sensors including its sensitivity to water vapor and the alignment of flow and gas signals for real time analysis. Recently, Phillips-Respironics has developed a novel oxygen sensor that utilizes luminescence technology for oxygen analysis. When combined with a differential-pressure transducer and an on-airway nondispersive infrared CO2 sensor, the result is a compact system suitable for real time breath-by-breath gas analysis. The system has been validated for use in a critical care environment with low respiratory flows of ±180 L/min. The purpose of this study was to determine the feasibility by modifying the existing breathing circuit to accommodate higher volumetric gas flows (±400 L/min) for exercise stress testing applications. Several variations of the prototype systems were constructed. To increase the flow, a differential pressure flow transducer was obtained from a commercially available system used for exercise testing. The gas analysis sensors were then inserted into the main lumen at a 45o angle so that the signal strength across the differential pressure drop was greater than 5 cm H2O at 400 L/min and produced minimal back pressure resistance. Characterization of the flow required the use of a flow coefficient, indexed by the Reynolds number, to adjust for head losses created by the differential pressure sensor. With the flow coefficient adjustments, the accuracy of the flow compared to the theoretical flow value was within ±3% or ±1 L. A propane combustion chamber that simulated oxygen consumption was used to validate the luminescence-quenching oxygen sensor. The fraction of expired oxygen was determined theoretically based on the complete combustion of propane and compared to the actual recorded valued. The sensor was found to be accurate to within 6% across the range of flow