5Non‐Ideal Data: Effects and Conditioning

5.1 Introduction to Non‐Ideal Data: Effects and Conditioning

Condition‐based data (CBD) contains feature data (FD) that forms signatures that are highly correlated to failure, damage, and degradation – especially fatigue damage due to the cumulative effects of stresses and strains induced by temperature, voltage, current, shock, vibration, and so on. Those signatures form curves that are not ideal, such as that shown in Figure 5.1; instead, CBD‐based signatures contain noise, they are distorted, and they change in response to that noise. That non‐ideality, when fault‐to‐failure progression (FFP) signatures are transformed into degradation progression signatures (DPS) and then into functional failure signatures (FFS), results in a non‐ideal transfer curve and errors in prognostic information. Those errors include the following: (i) an offset error between the time when degradation begins and the time of detection of the onset of degradation; and (ii) nonlinearity errors that reduce the accuracy of estimates of remaining useful life (RUL), state of health (SoH), and prognostic horizon (PH) – or end of life (EOL).

Graph of ratio vs. time displaying an ascending waveform for FFP: Experimental data intersected by an ascending curve for FFP: Simulated data.

Figure 5.1 Example of a non‐ideal FFP signature and an ideal representation of that signature.

5.1.1 Review of Chapter 4

Chapter 4 presented a set of seven signature models that resulted in FFP signatures having ideal, characteristic ...

Get Prognostics and Health Management now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.