Process gain answers a physical question: when you move the controller output by a certain amount, how far does the process variable eventually move in response? It is a property of the pipes, tanks, valves, and fluids themselves, not of the controller, and it is one of the three numbers a tuning method needs. Because most processes do not respond equally across their whole range, process gain often changes with operating point - and that changing gain is exactly what forces engineers to schedule different tuning on tanks and flow loops. This page keeps process gain distinct from the controller gain it is easily confused with.
Process Gain in one line: Process gain (Kp), also called steady-state or open-loop gain, is the ratio of the change in process variable to the change in controller output once both have settled - how much the PV moves per unit of output. It is a characteristic of the physical process, separate from controller gain, and when it varies across the operating range it drives the need for gain scheduling.
There are two gains in every loop and they are easy to conflate. Controller gain is a tuning parameter you choose - how aggressively the controller multiplies error into output. Process gain is a fact about the world you discover - how much the process variable moves for a given output change. You set the first; you measure the second. Confusing them leads to tuning that ignores the process, which is like choosing how hard to steer without knowing how sharply the car turns.
Process gain is calculated as the change in process variable divided by the change in controller output, after both have reached steady state, expressed in consistent units. If bumping a valve output up 10 percent eventually raises flow by 40 percent of span, the process gain is 4. A large process gain means a sensitive process where small output moves cause big PV swings; a small process gain means a sluggish process that barely responds. The units matter, so most engineers work in percent of span for both to keep the ratio dimensionless.
The two gains multiply together to set loop behavior, which is why they must be considered as a pair. The loop's overall aggressiveness is roughly the product of controller gain and process gain, so a high-process-gain loop needs a low controller gain to stay stable, and a low-process-gain loop can tolerate a high controller gain. Every tuning method is, at heart, a rule for picking controller gain to complement the measured process gain.
The complication is that process gain often is not a single number - it changes with where you are operating. Control valves are a common culprit: a valve with an equal-percentage characteristic gives a small flow change per unit of travel near closed and a large one near open, so the process gain seen by the controller can vary several-fold across the valve's range. Tuning that is comfortable at low flow can be unstable at high flow, or vice versa, purely because the process gain moved.
Tanks and vessels add their own nonlinearity through geometry. A horizontal cylindrical tank or a sphere changes level slowly per unit of volume near the ends and quickly in the middle, so the process gain of a level loop on such a vessel depends on how full it is. Blending, pH, and heat-transfer processes are famously nonlinear too, with pH being the extreme case where gain can change by orders of magnitude near neutrality. In all of these the single-number process gain is only a local snapshot.
When process gain varies enough to matter, one set of fixed tuning constants cannot serve the whole range, and the standard remedy is gain scheduling: storing several tuning sets and switching between them based on operating point, flow rate, or valve position. Choosing a valve characteristic that cancels the process nonlinearity - equal-percentage trim to offset an installed characteristic, for example - is the other classic tool, aiming for a loop whose overall gain stays roughly constant so one tuning fits everywhere.
Process gain is extracted from a bump test: put the loop in manual, step the output a known amount, wait for the process variable to settle, and divide the PV change by the output change. Doing this at several operating points, rather than just one, is what reveals whether the gain is constant or nonlinear. A single bump gives you a number; a bump at low, mid, and high range tells you whether that number can be trusted across the loop's territory.
A SCADA historian is what makes multi-point gain measurement practical on distributed assets. When a platform such as Merobix captures output and process variable together with reliable timestamps, an engineer can perform a bump remotely, read the settled values off the trend, and compute gain without a site visit - then repeat it at another operating point later. The historized record also lets you compare gain measured today against gain measured a year ago, which surfaces slow changes like valve wear or fouling.
Tracking process gain over time is a quiet but valuable diagnostic. A process gain that has drifted from its commissioning value often signals a mechanical change - a valve losing trim, a heat exchanger fouling, a line partially plugging - well before it shows up as an alarm. On remote oil and gas sites where nobody is watching the equipment directly, a shift in measured process gain, pulled from historized bump data, can be the earliest sign that a final element or the process itself is degrading.
Process gain is a physical property - how far the process variable moves per unit of controller output - that you measure from the process. Controller gain is a tuning parameter you choose that sets how aggressively the controller converts error into output. You measure process gain and select controller gain, and the two multiply to set overall loop aggressiveness.
Put the loop in manual, step the controller output by a known amount, wait for the process variable to reach a new steady state, then divide the change in PV by the change in output. Working in percent of span for both keeps the result dimensionless. For example, a 10 percent output step that produces a 30 percent PV change gives a process gain of 3.
When process gain changes with operating point - because of valve characteristics, tank geometry, or process chemistry - a single set of tuning constants that is stable in one region can be sluggish or unstable in another. Gain scheduling stores multiple tuning sets and switches between them based on flow, level, or valve position, so the loop stays well-behaved across its whole range.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.