Sensor fault diagnostics and prognostics using component model and time scale orthogonal expansions

A method of diagnosing sensor fault for an air handling system comprising the steps of:

• a) creating a component model by producing a series of measurement values and determining an expected value for a sensor measurement by producing a probability density function for the series of measurement values;
• b) detecting an actual measurement by the sensor;
• c) comparing the expected value to the actual value including comparing the probability density function to a normal distribution;
• d) determining the probability density functions of the outputs from a time-scale transformation,
• e) comparing the determined probability density function with a predefined probability density function when a sensor is healthy, and determining sensor health condition based on the results of the comparison.

A method of diagnosing a sensor fault condition for an air handling system comprising the steps of:

• a) creating a component model that represents operation of a heat exchanger and determining an expected value for a sensor measurement, wherein said operation of said heat exchanger is represented by the equations;
• Air
• ⁢
• ⁢
• side
• ⁢
• :
• ⁢
• ⁢
• Q
• =
• m
• .
• 1
• ⁢
• c
• p1
• ⁡
• (
• T
• 1
• ⁢
• in
• -
• T
• 1
• ⁢
• out
• )
• SHR
• Refrigerant
• ⁢
• ⁢
• side
• ⁢
• :
• ⁢
• ⁢
• Q
• =
• m
• .
• 2
• ⁡
• (
• h
• r1
• -
• h
• r2
• )
• where Q=the rate of heat transfer, {dot over (m)}1=mass flow rate of air, , {dot over (m)}2=mass flow rate of refrigerant, cp1=the specific heat of dry air, T=the temperature, SHR=the sensible heat ratio, hr1, hr2=specific enthalpies of refrigerant vapor at inlet and outlet of evaporator all in compatible units;
• b) detecting an actual measurement by the sensor;
• c) comparing the expected value to the actual value;
• and
• d) determining a fault in the sensor responsive to a difference between the expected value and the actual value greater than desired.

A method of diagnosing sensor fault for an air handling system comprising the steps of:

• a) creating a component model representing operation of a heat exchanger and determining an expected value for a sensor measurement, wherein a bias is estimated as the difference between a flow rate determined by a heat exchanger model and a flow rate determined by a compressor model;
• b) detecting an actual measurement by the sensor;
• c) comparing the expected value to the actual value;
• and
• d) determining a fault in the sensor responsive to a difference between the expected value and the actual value greater than desired.