CONTROL ALGORITHM IN INDUSTRIAL SYSTEM(PID CONTROLLER)

Algorithm means solution by steps. Control algorithms in industrial systems are mathematical formulas or sequences of instructions that manage and regulate various processes, with the aim of achieving a desired output. Specific tasks are reached by adjusting inputs based on feedback signals, thereby enhancing system performance, stability and efficiency. The system functions based on real-time feedback from the input and output signals.

In this section, I will discuss the P, PI and PID controllers.

P = Proportional gain (has 1 tuning parameter and depends on present error)

PI = Proportional and Integral gain (has 2 tuning parameters and an accumulation of past error)

PID = Proportional, Integral and Derivative gain (has 3 tuning parameters and prediction of the future error)

PI algorithm computes and transmits a controller output u(t) , signal every sample time t, to the final control element (e.g valves, variable speed pump). Computed controller output from PI is influenced by the controller tuning parameters and controller error, e(t)

u(t) = Kp e(t) + Ki ∫e(t) dt + Kd * de(t)/dt

Integral actions enables PI controller to eliminate offset, which is a major weakness of a P-only controller. PI is the most widely used algorithm in process control applications because it provide a balance of complexity and capability. At different times, the set-point is always the required speed or current which is set to be constant from a particular time, and error at any time can be calculated by subtracting the required speed and the controlled speed at that point in time.

Integral term, Ki, tends to increase the oscillatory or rolling behavior of process response. The PID controller is the best controller and widely used in industrial systems.

Integral windup refers to the situation in a PID controller where a large change in set-point occurs (say +ve change) and the integral terms accumulate a significant error during the rise (wind up), thus overshooting and continuing to increase as this accumulated error is unwound.

Solution to the integral windup includes;

• Initializing the controller integral to a desired value.

• Increasing the set-point in a suitable ramp.

• Disabling the integral function until the to-be-controlled process variable has entered the controllable region.

• Limiting the time period over which the integral error is calculated.

• Preventing integral term from accumulating above or below pre-determined bonds.