Graduation Semester and Year




Document Type


Degree Name

Master of Science in Electrical Engineering


Electrical Engineering

First Advisor

Kambiz Alavi


The project aims to develop a system to indicate leaks occurring in a Positive Airway Pressure (PAP) machine which sends a passage of air to the patients suffering from sleep apnea, through a hose and a nose mask. The system uses programming software to record and analyse the data obtained from the sensors and provides an interface for the user to select a method for notifying the time instant and location of leak occurrences.The system used an approach of using sensors for leak detection. Two pressure sensors and two audio sensors were selected out of which, one pressure sensor and audio sensor was installed on the mask side and the other pressure sensor and audio sensor was installed on the machine side. To install the sensors a connector was used for each side. The sensors were placed inside the connector. After the installation of the sensors, the signals from the sensors are acquired using a Data Acquisition device to read the analog signals and send it to the computer. A programming software was used to acquire the signals transmitted from the data acquisition device and perform necessary operations to analyse the data and detect leaks. The output of the sensors was too low for connecting with a data acquisition device, therefore circuits were designed to amplify the sensor signals. LabJack was chosen as a data acquisition device and LabVIEW software was used to acquire and analyse the signals transmitted by LabJack. Before analysing the signals, they were pre-processed for noise reduction and signal conditioning. The root mean square of the signal was found and used for analysis. The output voltages of the pressure sensor were converted to range of 4-20 as the pressure range of the PAP machine is 4-20 cm of water. Also, as the amplitude of the audio sensor did not change much for leaks, the area under the curve was found and used for leak detection. The leaks are classified based on the location as Mask and Machine leaks.The algorithm was only able to detect leaks of certain pressure ranges and it was not able to detect leaks of higher pressure range. Then the system was updated by replacing the sensors with an amplified and more sensitive one. The algorithm was also improvised by setting thresholds for each pressure range which increased the accuracy to 75%. The system was not still able to detect very fine leaks that occurs on the mask side. Therefore, the system was again updated with new programmable Microcontroller called Arduino and a new programming software called MATLAB. The Arduino was programmed to transmit the signals using serial port communication. A Moving Average filter with window size 500 was designed to reduce the noises in the system and the filtered signal values were again filtered to obtain a distinct signal for each set pressure with no leaks and for all intensities of leaks. The algorithm, instead of using the pressure sensor values separately, used the difference between the machine side and mask side pressure sensor values for detecting the leaks. The audio sensor on the mask side was eliminated and the audio sensor that was previously installed inside the connector was moved to outside for better response. An accuracy of 100 % was achieved in Laboratory testing. A Graphical User Interface was designed to start and stop the software and choose the notification method. Two notification methods were designed namely text message and alarm. Based on the user’s choice the corresponding notification method will be used to notify the user if the leak sustains longer than the period mentioned by the clinician.


Sleep apnea, PAP machine, Leaks


Electrical and Computer Engineering | Engineering


Degree granted by The University of Texas at Arlington