About the control algorithms
A control loop must meet a set of predefined quality requirements. These requirements depend on the control objectives, the process technology and the process itself. The behaviour of the closed-loop control must be examined and designed to meet the quality requirements.
The aim of the control is ensure that the system starts, operates (its output signals follow the predefined reference signals and suppresses the effect of the disturbances) and stops properly.
The starting and stopping of the system is usually controlled in an open loop, in consideration of logical conditions. The starting/stopping process usually does not depend on the output signal of the process.
The proper operation of the process is usually ensured by closed-loop control which manipulates the input signal of the process on the basis of the difference between the reference signal and the output signal. The basic connection is the serial control structure using negative feedback where the controller is serially connected to the process.
To control a process, open-loop and closed-loop control must be coordinated. Closed-loop control based on negative feedback can be completed with structural elements which can help improve the disturbance suppression properties.
In the control loop the difference between the reference signal and the measured output signal provides the input signal of the controller, and the output signal of the controller operates the actuator which provides the input signal for the system. The controller must be designed to ensure that the process satisfies the quality requirements set for the closed-loop control. The controller must be designed considering the process model and quality requirements.
If there are measurable signals (disturbances or internal signals) other than the output signal of the process, the control loop can be completed with further structural elements (e.g. disturbance compensation (feedforward), cascade control etc.). The more information is available, the more effective the control. If the process model is a state equation, it is advisable to involve state feedback. Control algorithms depend on the structure of the control and the model of the process.
It is important to tune properly the parameters of the controller in a given structure.
In an industrial context, PID controls are the most commonly used controls in single-loop feedback and cascade controls. PID controls provide the manipulating signal as the weighted sum of the error signal, its integral and its derivative. The controller relies on the input-output model of the process. Its parameters are often set by pole cancellation technique: the controller “drops” the poles that are unfavourable for the process and “brings in” poles with more favourable dynamics.
For systems described with a state model, we can design a closed-loop control which is based on the feedback of state variables, where the constants providing feedback from the states can be calculated by properly positioning the poles of the closed system or by optimising a quality criterion.