6.7 Control Variable

A control variable is the quantity within a system that is monitored and actively regulated by a control mechanism in order to maintain it at, or guide it toward, a desired value or trajectory. It is the primary output of interest in a control loop—the variable whose behavior defines whether the control system is achieving its objectives. The control variable is sometimes called the controlled variable, the process variable, or the regulated variable, depending on the engineering or scientific discipline and the context of application. Identifying the correct control variable is one of the most fundamental decisions in control system design, because the entire architecture of the feedback loop—what is measured, what is compared, what is adjusted—flows from this choice.

The distinction between the control variable and other system variables is important. A system may have many quantities that evolve over time, but only those designated as the control variable are actively regulated. Variables that are used to manipulate the control variable are the manipulated variables or control inputs; variables that perturb the control variable without being part of the control action are disturbance variables. In a room temperature control system, the room air temperature is the control variable; the power supplied to the heater is the manipulated variable; outdoor temperature, occupant density, and door openings are disturbance variables. The thermostat's entire function is to measure the control variable, compare it to the set point, and adjust the manipulated variable to drive the control variable toward the desired value despite disturbances.

The choice of control variable has profound implications for control system performance. The control variable must be measurable with sufficient accuracy and frequency. If direct measurement of the desired physical quantity is impossible or impractical, a surrogate variable that correlates closely with it may be used instead, but this introduces an additional source of error: the surrogate may deviate from the true variable of interest in conditions not anticipated during the control system design. In distillation column control, the product composition (the true control variable of interest) is often not directly measurable in real time, so temperature at a specific tray is used as a surrogate; this approach works well when the relationship between temperature and composition is stable, but fails when other factors (pressure fluctuations, feed composition changes) alter the temperature-composition relationship.

In multi-input multi-output (MIMO) systems, the control problem involves simultaneously regulating multiple control variables, each influenced by multiple manipulated variables. The interaction between channels—where a change in one manipulated variable affects multiple control variables—creates coupling that complicates controller design. The relative gain array (RGA) is a tool for analyzing these couplings and for selecting appropriate pairing of manipulated variables with control variables. The element λ_ij of the RGA measures the ratio of the open-loop gain to the closed-loop gain for the pair (manipulated variable j, control variable i), and values near 1 indicate that pairing is appropriate while large or negative values indicate problematic interactions.

In biological homeostasis, the control variables are the physiological quantities that must be maintained within narrow ranges for survival and normal function: blood glucose, arterial oxygen saturation, body core temperature, arterial pH, plasma sodium concentration, and many others. Each has dedicated sensor mechanisms (chemoreceptors, thermoreceptors, osmoreceptors) that measure the variable, dedicated comparison mechanisms that determine whether the current value falls within the acceptable range, and dedicated effector systems that respond to detected deviations. The physiological identity of a variable as a control variable is demonstrated by the existence of these dedicated regulatory systems: a quantity that has no sensors reporting it and no effectors responding to deviations from its normal range is not a biological control variable in this sense, regardless of its importance to physiology.

In organizational management, control variables are the performance metrics that the organization actively monitors and manages. Financial control variables include revenue, operating costs, profit margin, return on assets, and cash flow. Operational control variables include production volume, defect rates, on-time delivery percentage, and customer satisfaction scores. Each designated control variable has associated data collection processes (measurement), targets and thresholds (set points), reporting and review mechanisms (comparison), and corrective action processes (control response). The selection of organizational control variables determines what the organization attends to and responds to systematically, and what is left to informal attention or neglect.

The cascade of control variables in a hierarchical control system illustrates how the output variable of a lower-level loop can serve as the reference (set point) for that loop while being itself the manipulated variable for a higher-level loop. In a power plant, the steam pressure is a control variable in an inner loop regulated by fuel flow; the steam pressure set point is itself the manipulated variable of an outer loop that controls electrical output. This cascade architecture separates the control problem into layers of progressively faster and slower dynamics, allowing each loop to be independently tuned for its timescale. The cascade structure demonstrates that the role of a variable as "control variable" or "manipulated variable" is not an intrinsic property of the variable itself but a relational property defined by its position within the hierarchical control architecture.