A Monte Carlo Analysis of Condensation Risk from Sensor Error in Magnetic Resonance (MR) Cabinet

In a magnetic resonance imaging (MRI) cabinet cooling system, the temperature and humidity sensor (T/H sensor) is used to detect the temperature and relative humidity (RH) to worked out the specific humidity (SH) to judge whether anti-condensation system should be triggered. However, due to the unavoidable imprecision of the T/<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$H$</tex> sensor, the misestimation of actual SH may cause condensation damage or over-protection to the cooling system. This paper proposes a method to define a proper anti-condensation triggering SH threshold for MRI cabinet cooling system, which avoid the above mentioned two extreme conditions. The coolant temperature, ambient temperature and RH are assumed to follow normal distribution with the sensor-measured value as average <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mu)4$</tex> and sensor error as 3 times of standard deviation (3o), A Monte Carlo model was established to include these imprecisions and analyze the condensation risk probability without anti-condensation system triggered at a presupposed SH threshold. By selecting a threshold which makes the risk probability between 1 % and 2<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, the appropriate threshold can be determined. Through such a process, engineers can get an SH threshold for reference to avoid both condensation damage and over-protection due to T IH sensor imprecision in MRI cabinet.