Breath emulator for simulation and modelling of expired tidal breath carbon dioxide characteristics

In this work we describe a breath emulator system, used to simulate temporal characteristics of exhaled carbon dioxide (CO₂) concentration waveform versus time simulating how much CO2 is present at each phase of the human lung respiratory process. The system provides a method for testing capnometers incorporating fast response non-dispersive infrared (NDIR) CO₂ gas sensing devices - in a clinical setting, capnography devices assess ventilation which is the CO₂ movement in and out of the lungs. A mathematical model describing the waveform of the expired CO₂ characteristic and influence of CO₂ gas sensor noise factors and speed of response is presented and compared with measured and emulated data.

A range of emulated capnogram temporal waveforms indicative of normal and restricted respiratory function demonstrated. The system can provide controlled introduction of water vapour and/ or other gases, simulating the influence of water vapour in exhaled breath and presence of other gases in a clinical setting such as anaesthetic agents (eg N₂O). This enables influence of water vapour and/ or other gases to be assessed and modelled in the performance of CO₂ gas sensors incorporated into capnography systems. As such the breath emulator provides a means of controlled testing of capnometer CO₂ gas sensors in a non-clinical setting, allowing device optimisation before use in a medical environment.

The breath emulator uses a unique combination of mass flow controllers, needle valves and a fast acting switchable pneumatic solenoid valve (FASV), used to controllably emulate exhaled CO₂ temporal waveforms for normal and restricted respiratory function. Output data from the described emulator is compared with a mathematical model using a range of input parameters such as time constants associated with inhalation/ exhalation for different parts of the respiratory cycle and CO₂ concentration levels. Sensor noise performance is modelled, taking into account input parameters such as sampling period, sensor temperature, sensing light throughput and pathlength.

The system described here produces realistic human capnographic waveforms and has the capability to emulate various waveforms associated with chronic respiratory diseases and early stage detection of exacerbations. The system has the capability of diagnosing medical conditions through analysis of CO₂ waveforms. Demonstrated in this work the emulator has been used to test NDIR gas sensor technology deployed in capnometer devices prior to formal clinical trialling.