CO2-based demand control ventilation (DCV) maintains the CO2 concentration in the rooms within an appropriate range by adjusting supply air flowrate.
If the CO2 sensor is faulty, indoor air quality and energy saving cannot be guaranteed.
An automatic CO2 sensor fault detection, diagnosis and self-correction method is proposed in this study. The basic idea is to get benchmark values of CO2 sensors by CO2-uniform indoor environments in buildings. One way is to recycle air without any outdoor air for 1-2 hours, i.e. 100% return air ventilation.
All CO2 sensors should have a same reading in the end. Another way is to supply fresh air into buildings without recycling air for about 1-2 hours, i.e. full outdoor air ventilation. Readings of all the CO2 sensors should equal to the CO2 concentration of the ambient air in the end. Faulty sensors are found if their readings are different from benchmark values.

Demonstrations for these methods are made by simulating of a section of a school building. Results show that the proposed method is effective to detect, diagnose and remove soft CO2 sensor faults.