What Is Fault Accommodation?

Fault accommodation is a control engineering strategy in which a system automatically adjusts its operation to maintain acceptable performance after a fault has occurred in one of its components. Rather than shutting down or failing abruptly when a sensor, actuator, or subsystem degrades, a fault-accommodating system modifies its control law, switches to a backup configuration, or reweights available inputs and outputs to compensate for the degraded element. The goal is to preserve stability and meet mission objectives, even if performance levels are somewhat reduced from nominal.

Fault accommodation belongs to the broader category of fault-tolerant control (FTC), which encompasses any strategy designed to keep a closed-loop system functional in the presence of component failures. Its closest companion concept, fault adaptive controls, generalizes accommodation by employing adaptive mechanisms that can respond to a wider range of fault types and severities without requiring explicit pre-designed reconfiguration logic.

Fault Detection and Diagnosis as a Prerequisite

Effective fault accommodation depends on first knowing that a fault has occurred and characterizing its nature. Fault detection and isolation (FDI) algorithms monitor system outputs, residuals between measured and model-predicted values, or statistical indicators to identify when a deviation from normal behavior has arisen. Once a fault is detected, a diagnosis step estimates its location, type, and magnitude. Reconfigurable control system design for fault diagnosis and accommodation describes how these two phases, detection-isolation and controller redesign, are coupled in a unified architecture. The quality of accommodation depends directly on the speed and accuracy of the FDI step: a slow or ambiguous diagnosis delays reconfiguration and may allow the system to drift into an unsafe region before corrective action is applied.

Accommodation Strategies

Once a fault is characterized, accommodation is achieved through one of several strategies. Re-tuning adjusts controller gains or setpoints to account for changed plant dynamics without altering the control structure. Virtual sensing replaces the signal from a faulty sensor with a software estimate derived from a dynamic model and healthy measurements, a technique called analytical redundancy. Control reconfiguration, the most general approach, modifies the control law itself, redistributing authority across remaining actuators or selecting an entirely different control mode. In fault-tolerant control for industrial machines, reconfigurable and adaptive solutions combine model-based diagnosis with online controller redesign, allowing the system to handle unanticipated failure modes that were not explicitly enumerated during design. A tutorial introduction to reconfigurable fault-tolerant control surveys these strategies and their trade-offs in terms of computational cost, coverage of fault types, and time to achieve stable reconfiguration. Passive accommodation strategies, by contrast, build robustness into the nominal controller through conservative gain margins and redundant actuator allocation, so that certain faults cause no detectible performance loss without any active response.

Design Considerations

Designing a fault accommodation system requires specifying which fault scenarios the system must handle, what performance degradation is acceptable, and how quickly accommodation must occur. Formal requirements are often expressed as degraded stability margins or bounded tracking error in the presence of listed fault types. Verification typically uses simulation of fault injection scenarios combined with hardware-in-the-loop testing. Safety-critical applications impose additional requirements on the accommodation system itself: the reconfiguration logic must not introduce instability, and its own failure modes must be analyzed. Standards bodies including SAE and EUROCAE address these requirements in certification frameworks for aviation and automotive safety.

Applications

Fault accommodation has applications in a wide range of engineering fields, including:

• Aerospace, where flight control systems must maintain controllability after actuator or sensor failures
• Industrial process control, where continuous plants cannot tolerate unplanned shutdowns
• Autonomous vehicles, where onboard systems must compensate for sensor degradation in real time
• Power systems, where protection schemes reconfigure network topology after equipment faults
• Robotic systems, where joint or drive failures require redistribution of motion authority