6 Regulation
6.1 Regulation Control

Higher level for regulation is automatic control. Means that the system is controlled so that one or more physical variables are maintained at the prescribed parameters. An example of such a system may be gas furnace - controlled system, which performs preheating of material for reasons of a surface treatment (e.g. tempering). Material supplied inside the oven must be heated to a specified temperature and the temperature of the furnace, the regulatory circuits (control and regulation) to maintain a certain period of time (set point). Temperature must be measured and its size is controlled by control valve (actuator) supply of fuel gas. The system is controlled by one or more of the parameters measured. These may be any physical quantity: temperature, pressure, speed, power, voltage, etc. Measuring unit processes the measured value to the appropriate signal and passes it to the controller. Since this is a transfer of information from the system, this branch is called feedback. Also entering into the controller set point (value). It is the size of the regulated parameter. The difference between value and reference signals feedback control deviation occurs. Control deviation signal enters the control block, the size of which creates the appropriate control input to the actuator. Their activities affect the actuator system and its parameters.

image
Control regulation scheme

Control

Control is a process, using a feedback to reach desired aim (feedback control). The aim of the control is to reach and ensure desired value of controlled (output) value (e.g. room temperature, tank level) or the desired time behaviour (e.g. temperature behaviour according to weekly plan or a temperature in room according to specifications). The desired value must be assured not only after desired value change, but under disturbances, acting to a system, as well. The disturbances typically do have an unpredictable characteristic, e.g. thermal loss or increase in heated room (outdoor temperature change, window opening, draught, wall and room insolation, presence of persons or powered electric equipment).

Feedback

Principal schematic of feedback control system is shown on the Figure above. The input of the whole system is a desired value (w) and an output is the actual value (y). The subtraction element evaluates an error e = w – y, what is input to a controller. The controller processes the error and outputs a control variable, which acts via actuators to the controlled system (plant). The controller tries to minimize the error, for the actual value y to approach the desired w.

PID controllers

PID is the most used type of controller. The common property of ordinary P, PI and PID controllers is linearity. In case of proportional controller (P), the control variable is directly proportional to the error.

The control variable of proportional-integral controller (PI) is a sum of two components – the proportional one (which is, as well as in case of pure P controller, directly proportional to the error), and an integral one, which is proportional to accumulated value of the error, i.e. to its integral. Integral action is able to reach zero error in some cases, where it is impossible with pure proportional controller.

The output (control variable) of proportional-integral-derivative controller (PID) contains an additional derivative action. It has "anticipating" behaviour and brings faster response to sudden changes. Its disadvantage is that it amplifies high frequency noise, present in measurement, what may cause random, erratic operation of the system.

Until recently, PID controllers were implemented as analogue circuits, usually based on operational amplifiers. Controllers now are usually implemented in software. The software can run on a microcontroller, digital signal processor or a PLC in case of industrial application, or an ordinary personal computer.

Evaluates mathematical expression

uk = p·ek + i·Σek + d·Δek

Integral of the error is replaced by sequential sum of individual error value samples at each step (Σek = Σek-1 + ek). Derivation is replaced by a backward difference, i.e. the difference between actual and previous error sample Δek = ek – ek-1.