Faults incipient

Watanahe, K. and D.M. Himmelhlau, Incipient Fault Diagnosis of Nonlinear Processes with Multiple Causes of Faults, Chemical Engineeiing... 

A fault may interfere with the effectiveness or the func tioning of the unit (Watanabe, K., and D.M. Himmelblau, Incipient Fault Diagnosis of Nonhnear Processes with Multiple Causes of Faults, Chemical Engineering Science, 39(3), 1984, 491-508). The first question addresses the effectiveness. The second two address the functioning. Fault detection is a unit monitoring activity, done automatically or periodically, to determine whether the unit operation has changed. 

The NPRDS is an industry-wide system for monitoring the performance of selected systems and components at U.S. commercial nuclear power plants. Information in NPRDS is derived from a standardized format input report prepared by U.S. nuclear plant licensees. The plants are as)ced to submit failure reports on catastrophic events and degraded failures within the defined reportable scope reporting of incipient events is optional. Command faults are not reportable unless they malce an entire system unavailable. In addition, the plants are as)ced to file component engineering reports on all components within the selected systems and reportable scope. These reports contain detailed design data, operating characteristics, and performance data on the selected systems and components (over 3000 components, from approximately 30 systems, per unit). The selected systems are primarily safety systems. 

X. Zhang, M.M. Polycarpou, and T. Parisini. A robust detection and isolation scheme for abrupt and incipient faults in nonlinear systems. IEEE Transactions on Automatic Control, 47 576-593, 2002. 

Root-cause detection Incipient fault detection Problem diagnosis Failure prognosis Post mortem... 

Finally, with regard to condition based maintenance (CBM) of engineering systems it is important to detect the initiation of incipient faults, to assess the current health of a system and to predict the remaining useful life (RUL) of a component that has become faulty. As bond graph modelling enables the systematic generation of ARRs for FDI, it can also support failure prognosis. 

There is an abrupt fault if at some time instant a system component suddenly turns from normal into abnormal operation. A fault is of progressive nature if instead the behaviour of a component gradually deviates from that of normal operation. If the behaviour of a system component does not match the desired behaviour but is still within given bounds then this is an indication for an incipient fault. As incipient faults develop slowly, are small in comparison to abmpt faults and may be obscured by disturbances, they are rather difficult to detect. They are usually due to wear and tear, and to ageing in normal operation and, although the system behaviour is still acceptable, they should be detected so that an identified component causing an incipient fault can be replaced before the component s behaviour turns into a non-acceptable malfunction that can cause a failure. 

The method is particulary suited for the detection of incipient faults as it can pinpoint small parameter changes. In Chap. 6, least squares parameter estimation is considered with to regard to the isolation of multiple parametric faults. 

For these reasons, in the following, it is assumed that no parametric faults happen during system mode identification and that an initial system mode is known. System mode identification in the presence of faults is more difficult. In [4], Arogeti et al. present an advanced method for this more general case that categorises ARRs into different types and provides a refined set of fault candidates to the fault parameter estimation procedure. Multiple fault detection, isolation and identification for hybrid systems with no available information on the nature of faults (abrupt or incipient) and on system mode changes has been recently addressed in [5],... 

By using thresholds and a decision procedure, fault diagnosis can provide an indication of the onset of an incipient fault that may lead to a failure. If a model of the anticipated degradation of parameters is available, the time evolution of monitored parameters or features can be extrapolated. The time from the starting point of a fault until the projected parameter value intersects with a user defined alarm threshold is then an estimation of the remaining useful life. The prediction of the RUL is affected by uncertainties in the monitored parameters, by the choice of the degradation model, uncertainties in its parameters and uncertainties in the failure alarm thresholds. As... 

Let 0 denote the vector of component parameters i,..., 0, let y(f) be the vector of known measured outputs of the real system, (f) the vector of inputs into the system and into the nominal model of the system and wi (f) the vector of all switch states, i.e. m t) denotes the system mode. Assume that fault diagnosis has detected and isolated 1 as a single incipient fault starting at time instant to. Then... 

Degradation models for system components may be determined by accelerated life tests and then used after online fault diagnosis has identified the initiation of an incipient fault in order to assess the health of a system and to predict the RUE of the component with an incipient fault [8]. 

Given a failure alarm threshold and the starting point to of an incipient fault, then the solution of the degradation model can be used to determine the time as of the starting point of the incipient fault until the degradation trend of the parameter intersects... 

The DBG of the network in Fig. 9.4 accounts for a degradation of the voltage supply indicated by the efficiency and an incipient fault in resistance R2. 

Figure9.7a displays the time history of the capacitor voltages Mi, M2 and their degradated waveforms u t),U2 t) due to the incipient faults in the voltage supply and in resistance R2-...

In the faulty system, the supply voltage is not stable but decreases as of r = 5 s and so do the capacitor voltages. The effect of the incipient faults on the resistor... 

Assume that online monitoring and fault detection produce a coherence vector c = [1 0] in the time interval 0 < < 10 s when the circuit is in mode b = 0. A comparison of the coherence vector with the FSM in Table 9.2 reveals efficiency p and resistance Ri as possible fault candidates. Parameter estimation can identify p as a true fault. However, the incipient fault in R2 starting simultaneously with the fault in /3 at = 5 s cannot be detected. Residual ri is not sensitive to a fault in R2 and the term /(/ on + Riit)) in ARR2 and ARR3 is cancelled out by the switch state b... 

For systems described by a hybrid model at least some ARRs are mode-dependent. As a result, incipient faults may start when the system is in a mode that does not enable an ARR based approach to detect the fault because its contribution to an ARR or to a set of ARRs is cancelled out by switch states being equal to zero in that mode. An approach adopted from [6] and illustrated by a simple passive network with one switch is to allow for a delay time in fault isolation at least until the next mode change has happened so that faults that cannot be detected in the current mode can be detected and isolated after a mode change. 

Prognosis Failure prognosis means the ability of an early detection and isolation of incipient faults that may lead to a component failure, to determine the progression of the fault and to predict the remaining useful life (RUL), i.e. the time to failure given the current state of a system. 

---