PID Control Loops in Fusion Systems

The proportional-integral-derivative loop is the workhorse regulator behind coil currents, temperatures, flows, and pressures throughout the plant.

The three terms

A PID controller computes an actuator command from the error between a measured value and its set point. The proportional term reacts to the present error, the integral term accumulates past error to remove steady offset, and the derivative term anticipates by reacting to the error's rate of change. Their gains are tuned to give a fast, stable response.

def pid_step(err, integ, prev_err, dt, kp, ki, kd):
    integ += err * dt
    deriv = (err - prev_err) / dt
    out = kp*err + ki*integ + kd*deriv
    return out, integ, err

Where it is used

PID loops regulate power-supply currents, cooling-loop temperatures, vacuum-relevant pressures, gas-valve flows, and many balance-of-plant utilities. Each loop treats its actuator and sensor as a simple pairing. The strength of PID is that it needs little model of the system, only tuning, which makes it the default for the plant's many single-variable loops.

Tuning and limits

Gains are tuned for the loop's dynamics: too much gain oscillates, too little is sluggish. Practical loops add anti-windup so the integral term does not accumulate while the actuator is saturated, and output clamps to respect hardware limits. Derivative action is often filtered because it amplifies noise. These refinements make a textbook PID robust in a real plant.

• Proportional reacts, integral removes offset, derivative anticipates
• Default regulator for single-variable loops
• Anti-windup handles actuator saturation
• Filter derivative to reject noise

When PID is not enough

Strongly coupled or fast-unstable problems, like plasma shape and vertical stabilization, exceed what independent PID loops handle well and use multivariable or model-based control instead. But even those systems rest on fast inner current loops that are often PID. In the Kronos control architecture, described as a design and simulation model, PID regulators form the base layer beneath the specialized plasma controllers.

Most of a plant's regulation is PID doing its steady, dependable work out of sight.