Method and apparatus for trending and predicting the health of a component

A method executed in a computer system for determining an health indicator of a component at a subsequent time comprising:

• determining a first health indicator of said component at a time, n, in accordance with at least one corresponding condition indicator;
• using a three state Kalman filter to determine a second health indicator of said component at a time subsequent to time n;
• and
• estimating said first health indicator using a hypothesis determination technique, wherein said at least one condition indicator is a normalized condition indicator.

A method executed in a computer system for determining an health indicator of a component at a subsequent time comprising:

• determining a first health indicator of said component at a time, n, in accordance with at least one corresponding condition indicator;
• and
• using a three state Kalman filter to determine a second health indicator of said component at a time subsequent to time n, wherein H = [ 1 0 0 ] Φ = [ 1 dt dt 2 2 0 1 dt 0 0 1 ] Q = 2 ⁢ ⁢ σ 2 ⁢ t _ ⁡ [ dt 3 2 dt 2 2 t _ dt 2 2 dt 1 t _ 1 1 t _ ] X = [ HI_est H . ⁢ I H ¨ ⁢ I ]
• in which:
• σ is a power spectral density,
• R is a measurement error,
• P is a covariance,
• Q is a plant noise,
• H is a measurement matrix,
• K is a Kalman gain and
• Φ is state transition matrix, and
• Xt|t−1=ΦXt−1|t−1 (Equation T1)
• Pt|t−1=ΦPt−1|t−1ΦT+Q (Equation T2)
• K=Pt|t−1HT(HPt|t−1HT+R) (Equation T3)
• Pt|t=(I−KH)Pt|t−1 (Equation T4)
• Xt|t=Xt|t−1+K(HI−HXt|t−1 (Equation T5)
• and the method further comprising:
• using the above Equations T1 through T5 to determine a value Hi_est which is a best estimate of a value HI at a time in the future.

A computer program product for determining an health indicator of a component at a subsequent time comprising machine executable code for:

• determining a first health indicator of said component at a time, n, in accordance with at least one corresponding condition indicator;
• and
• using a three state Kalman filter to determine a second health indicator of said component at a time subsequent to time n;
• and
• estimating said first health indicator using a hypothesis determination technique, wherein said at least one condition indicator is a normalized condition indicator.

A computer program product for determining an health indicator of a component at a subsequent time comprising machine executable code for:

• determining a first health indicator of said component at a time, n, in accordance with at least one corresponding condition indicator;
• and
• using a three state Kalman filter to determine a second health indicator of said component at a time subsequent to time n, wherein H = [ 1 0 0 ] Φ = [ 1 dt dt 2 2 0 1 dt 0 0 1 ] Q = 2 ⁢ ⁢ σ 2 ⁢ t _ ⁡ [ dt 3 2 dt 2 2 t _ dt 2 2 dt 1 t _ 1 1 t _ ] X = [ HI_est H . ⁢ I H ¨ ⁢ I ]
• in which:
• σ is a power spectral density,
• R is a measurement error,
• P is a covariance,
• Q is a plant noise,
• H is a measurement matrix,
• K is a Kalman gain and
• Φ is state transition matrix, and
• Xt|t−1=ΦXt−1|t−1 (Equation T1)
• Pt|t−1=ΦPt−1|t−1ΦT+Q (Equation T2)
• K=Pt|t−1HT(HPt|t−1HT+R) (Equation T3)
• Pt|t=(I−KH)Pt|t−1 (Equation T4)
• Xt|t=Xt|t−1+K(HI−HXt|t−1) (Equation T5)
• and the computer program product further comprises machine executable code for:
• determining, using above Equations T1 through T5, a value Hi_est which is a best estimate of a value HI at a time in the future.