The Continues Positive Airway Pressure (CPAP) therapy is the most popular noninvasive therapy used to treat Obstructive Sleep Apnoea symptoms (OSA). Recent sleep clinic research requires more CPAP functions during patient s sleep periods. It would be an advantage to operate the CPAP system at a lower pressure whilst monitoring the patient s air flow, and raise the pressure as the OSA is detected, rather than maintaining a fixed higher minimum pressure in order to reduce OSA symptoms. In order to detect OSA, an air flow sensor is considered ideally suited to assemble within the CPAP system and the characteristics of this type of sensors are investigated by creating and validating a mathematical model.

Before determining the most suitable type of air flow sensor, several common methods of air flow measurement were investigated. Ultrasonic and hot wire air flow sensors were considered to be the most suitable sensors to be implemented. After setting up the simulation model and conducting experiments, some limitations of ultrasonic air flow sensor were found, so the hot wire sensor was selected for further investigation.

Comparing the performances of three types of hot wire sensor, Constant Voltage (CV), Constant Current (CC) and Constant Temperature (CT), the CT hot wire sensor was found to be the most suitable type for CPAP systems. In order to achieve a precise measurement, different locations for the air flow sensor within CPAP system were investigated and the blower inlet was found to be the best location for air flow sensor.

The simulation CT model hot wire sensor was developed. The experimental results were compared to the simulation model results. Comparison of the two results showed less than 1.5 % error when the air flow velocity exceeded 1.0 m/s (Average air speed is 5.0 m/s from the CPAP Blower).

Environmental conditions of temperature, humidity and wire property effects were analysed by processing the simulation model under different conditions. The following conclusions were drawn out: The output current to the hot wire air flow sensor rise significantly as the hot wire diameter increases, however, the sensor response time increases significantly. It was found that platinum wire of a diameter 0.02 mm can meet the response time requirement for diagnosis of sleep disorders.

Various wire lengths produce no significant effect on both aspects of output wire current and response time. Therefore wire length can be designed to fit the inlet channel of the CPAP blower.

Higher wire temperature causes higher wire current flow at the same air flow velocity, however, lower wire temperatures achieve a faster response speed. The temperature range of 50 to 100oC over ambient was found to be an optional setting.

To achieve the desired level of accuracy a temperature compensation circuit was required for the hot wire air flow sensor. Otherwise it was necessary to keep the sensor in a stable room temperature environment to gain the same result. The percentage error of ambient temperature was about 0.65% per oC.