A context-appropriate oxygen concentrator for low-income countriesFollow project
Typical oxygen concentrators use a pressure swing adsorption cycle (PSA) to scrub nitrogen out of a compressed air supply, leaving high concentration oxygen, greater than 85%, as an output. The process involves two molecular sieve beds, each containing zeolite, and a four-step Skarstrom cycle as shown below. The highly porous zeolite particles have a large specific surface area and a selectivity for adsorbing nitrogen molecules in a purely kinetic process. Once a sieve bed is pressurised, kinetic adsorption takes place until the adsorbent surface is fully utilised, it is then depressurised and a backflow of oxygen from the other sieve bed is used to release remaining nitrogen from the zeolite to the atmosphere.
The concentration of oxygen produced by many commercially available devices has been shown to decrease dramatically under harsh environmental conditions. A combination of high temperature, relative humidity and dust concentration produces these performance drops by affecting the chemical reactions, blocking the filters and damaging the electronic components.
Limited spare parts supply chains affect the regularity and effectiveness of maintenance on oxygen concentrators. The aim of this project is to develop solutions that reduce the maintenance requirements whilst improving the ease of servicing these devices.
As shown below, an Arduino Mega is used to control solenoid valves to direct the gas flows through the device. The microcontroller also receives signals from sensors monitoring the cyclic process.
Pressure sensors such as the NXP MPX5700DP are used to monitor the pressure at various points within the device. Monitoring the oxygen outlet pressure allows a technician to understand whether leaks have formed in the system due to joint failures, or whether blockages are causing increased pressure drops. Monitoring the pressure at each molecular sieve bed would highlight when the compressor is taking longer to pressurise each bed, allowing the cycle time to be automatically adjusted by the Arduino microcontroller in order to maintain the most efficient operating conditions.
Temperature sensors such as the EPCOS M703 NTC probe can be mounted to the compressor piston sleeves, enabling the operating temperature of the sleeves to be monitored. If these are running too hot, this can cause premature failure of the sleeves which need replacing by a technician. Being able to monitor this and provide alerts before the sleeves fail means that action can be taken to ensure the compressor cooling system is functioning properly.