Abstract: A fault detection and diagnosis method is presen- ted, which based on analytical redundancy and statistics, for detecting and evaluating the temperature sensors with fixed bias in HVAC water systems. Based on energy conservation, analytical redundancy models of measurements and measure- ment residuals of each selected control volume are set up, and then the FDD problem could be changed into solving the minimization problem by using of statistic theory. The equations for FDD are set up and solved to obtain the fixed bias of measurements of temperature sensors. It is discussed to select data samples for FDD as well as reliability in this paper. The presented FDD method is tested and validated by simulation of chilling water systems. The results show that this method could be used to detect and evaluate the fault of fixed bias of temperature sensor caused by zero shifts.

Key words: FDD, Fixed bias, HVAC, Temperature sensors, Water systems