(英) |
To measure carbon dioxide concentration in many rooms, a large amount of CO$_2$ sensor is required.
The CO$_2$ sensor is divided into two types, optical type and contact type, the optical type is a non-dispersive infrared type (hereinafter referred to as NDIR type), the photoacoustic type, the contact type is a solid electrolyte type, an electrochemical type, a semiconductor There is an expression.
The NDIR method which is common in CO$_2$ sensors is very accurate because it senses only specific gases, but the operation cost is high because the unit prices are as high as 15,000 yen for cheap ones.
On the other hand, if the semiconductor type CO$_2$ sensor is used, the operation cost is low because it is 700 yen for a cheap item and the unit price is low.
Therefore, in this research, in order to measure with a low cost and low accuracy CO$_2$ sensor, we attempted to correct the CO$_2$ sensor with low precision based on the data acquired with the highly accurate CO$_2$ sensor.
The accuracy of the correction method using the multiple regression equation calculated by multiple regression analysis and the correction method using the prediction by the linear regression model were compared and examined. |