Glossary of Control Engineering Terms - F

Fault Tolerant: The ability to withstand a considerable degree of error without crashing or going unstable. This may involve copying or mirroring data and having extra capacity. It is often achieved by using several separate processors and/or sensors and actuators, each monitoring the others and operating a voting system, so that if one fails, the others shut it down, call for human help and carry on operations. Some control techniques can be designed to have fault tolerance.

Feed-forward Compensation: If a disturbance from another measurement or observer, then you can compensate for the disturbance by applying appropriate controller output. This usually requires an accurate model of the process and is used in conjunction with closed loop control to eliminate errors in the model.

Fieldbus: Fieldbus is a generic term used to describe a common communications protocol for control systems and/or field instruments.

First-Order System: A dynamic, linear system that has one pole, usually characterised by a gain and a time constant. A first order system is typical of many physical processes, such as how levels change in gravity fed systems. A first order system can never exhibit overshoot. Also referred to as a Lag System.

Fourier Transform: this coverts a time response to frequency response. It is used within signal processing, but the closely related Laplace Transform is used for linear systems due to its better properties.

Frequency: Defines number of cycles during a time period. Hertz (Hz = 1 cycle per second) and radians per second (rad/s, 1Hz = 6.283 rad/s) are units of frequency.

Frequency Response: how the output of a system varies with frequency. Typically at slow frequencies a system will respond with unity gain and zero phase shift, at faster frequencies the dynamics of a system means that it won't be able to keep up so well resulting in attenuation and phase shift. Frequency responses and time responses are related through the Fourier Transform and Laplace Transform.

Fuzzy Logic: Fuzzy logic was formalised by Lofti Zadeh to overcome the inadequacies of conventional logic for analysing real world systems. Rather than using conventional logic which states that an attribute either belongs to a set or does not (i.e. something is hot or cold), it assigns a degree of membership of an attribute to a particular set (i.e. something is "very" hot and "barely" cold). These linguistic variables can be combined in various ways to replicate complex behaviours (e.g. if car is approaching a large hill, change gear if rpm is too low) and control strategies.