NASA Logo, National Aeronautics and Space Administration

Fault Modes

A number of fault modes have been identified so far for implementation on the testbed (Table 2). The criteria for their selection include relevance to a variety of aerospace vehicles (not just rovers), feasibility of implementation, and progression time from fault to failure. The last criterion is important because if the progression time is too brief (e.g. microseconds), then likely no useful action can be taken in the prognostic context to predict the remaining useful life of the component and remedy the situation. On the other hand, if the fault-to-failure progression time is measured in years, then running experiments on those fault modes may become impractical. Faults in both of the above categories could still be handled by diagnostic systems, however. Out of the fault modes described in Table 2, a few were selected for the initial phase of the project. The methods for their injection on the K11 are covered in more detail next. The methods for modeling progression of these faults in the simulator are described in Section 3.

Mechanical Jam and Motor Windings Deterioration

The first fault mode selected for implementation is a mechanical jam on the motor axle which leads to increased current, overheating of motor windings, deterioration of their insulation, and eventual failure of the motor due to a short in the motor windings. To maintain realism, a performance region for the motor is chosen (using manufacturer’s specifications) where a healthy motor would have no problems keeping up with either speed or load requirements. In the presence of increased friction, however, the amount of current needed to satisfy the same speed and load demands is higher, leading to overheating. Unless speed and/or load are reduced or duty cycle (the proportion of time the motor is on versus duration of cool-down intervals) is adjusted, the heat build-up will eventually destroy the insulation of the motor windings and lead to motor failure. This fault mode was first implemented in the simulator and its model verified using experimental data collected on smaller-sized motors that were run to failure under similar conditions (please see section 3.2, Motor Modeling). A hardware fault injection using a mechanical brake on one of the rover motors will be implemented next. The rover motor will not be run to complete failure initially; instead the simulator model parameters and prognostic algorithms will be validated in experiments stopping short of creating permanent damage. Eventually, experiments that will take motors all the way to failure will be performed.

Parasitic Load

A parasitic electrical load will be injected on the main power distribution line via a remotely controlled rheostat. The rheostat can be set for resistance from 0 to 100 Ohms and can dissipate up to 600 Watts of power. The rheostat will simulate a situation where, for example, an accessory motor is continuously engaged due to a failed limit microswitch.

Electronics Faults

The systems on the K11 provide several opportunities for fault injection in electronics subsystems. Power electronics in the motor drivers allow fault injection in power-switching devices such as Metal–Oxide–Semiconductor Field-Effect Transistors (MOSFETs), Insulated Gate Bipolar Transistors (IGBTs) and electrolytic capacitors used for voltage filtering. These devices have a key role in providing current to the motors, but are known for relatively high failure rates. Fault injection will also be implemented on the power switches of the motor winding H-bridges, where current will be routed to degraded power transistors during rover operation. In addition, some of the symptoms of power transistors failing will be replicated programmatically by varying the gate voltage. The premise of the fault injection in the H-bridge transistor is that it will diminish the performance of a motor winding, reducing torque and altering motor performance characteristics, making control difficult.

Efforts on accelerated aging of IGBTs and power MOSFETs are presented in (Celaya, Saxena, Wysocki, Saha, & Goebel, 2010). Accelerated aging methodologies for electrolytic capacitors under nominal loading and environmental conditions are presented in (Kulkarni, Biswas, Koutsoukos, Celaya, & Goebel, 2010); methodologies for accelerated aging via electrical overstress are presented in (Kulkarni, Biswas, Celaya, & Goebel, 2011). MOSFETs, IGBTs, and electrolytic capacitors at various levels of degradation will be used to inject component-level electronic faults, with some of the faults expected to have a cascading effect on other electronic and/or mechanical subsystems.

Battery Capacity Degradation

As the rover batteries go through charge/discharge cycles, their capacity to hold charge will diminish. The degradation rate will depend on several factors such as imposed loads, environmental conditions, and charge procedures. For example, Li-Ion chemistry batteries undergo higher rates of capacity fade with higher current draw and operational temperatures. Even at rest, this type of battery has chemical processes occurring that have long-term effects - for instance, latent self-discharge and transient recovery during relaxation. The depth-of-discharge (DoD) and even the storage temperature have major influences on the overall life of the battery as well. There is no specific mechanism required for injecting this fault – the batteries will age naturally in the course of rover operations. Some experiments will, however, utilize battery cells aged to a desired point in their life cycle on the battery aging test stand (Saha & Goebel, 2009)

Remaining Battery Charge Tracking

While not being, in the strict sense, a fault, tracking the remaining battery charge will be one of the main tasks of the prognostic system. End of charge is an end-of-life criterion, so the remaining charge estimate is expected to be a factor in most of actions undertaken by PDM software. Most battery-powered devices have some form of battery state-of-charge (SOC) monitoring onboard. This is mostly based on Coulomb counting, i.e. integrating the current drawn over time, divided by the rated capacity of the battery. The definition used in this work is the following: SoC= 1-(∫_(t=0)^(t|V=V_cutoff)▒I(t)dt)/"Capacity of Current Cycle" ×100% It should be noted that both the numerator and denominator of the fraction are predictions, not the actual measurements: battery voltage prediction for the former and capacity prediction for the latter. Further details are discussed in (Saha and Goebel 2009).

First Gov logo
NASA Logo - nasa.gov