Cascade Control: Loop Interaction and Design

Core Idea

Cascade control uses an inner fast loop to control an intermediate variable and an outer slow loop to control the final output. Inner loop reduces effective disturbance entering outer loop, improving disturbance rejection. Design is hierarchical: inner loop must be stable and fast, then outer loop is designed treating inner loop as part of the plant.

Explainer

From your study of feedback control fundamentals, you know that a single feedback loop compares the output to a setpoint and adjusts the manipulated variable to reduce the error. This works well when disturbances enter near the process output — the sensor detects them quickly. But many real processes have disturbances that enter early in the process chain, far upstream of the output sensor. By the time the output deviates and the single-loop controller reacts, the disturbance has propagated through the entire plant. Cascade control addresses this by adding a second, faster loop that intercepts disturbances before they reach the primary output.

The architecture has two nested loops. The inner loop (also called the secondary loop) measures an intermediate process variable — one that is closer to where disturbances typically enter and responds faster than the final output. The inner controller acts quickly to regulate this intermediate variable. The outer loop (primary loop) measures the final controlled variable and generates a setpoint for the inner loop, rather than directly commanding the actuator. The outer controller essentially says "make the intermediate variable equal to this value," and the inner loop executes that command rapidly. From the outer loop's perspective, the inner loop and the physical path from intermediate variable to output become a faster, better-behaved "plant."

The design is deliberately hierarchical and sequential. The inner loop is designed first: it must be stable and significantly faster than the outer loop — a rule of thumb is that the inner loop's closed-loop bandwidth should be at least 3–5 times faster than the outer. If this separation of timescales is not respected, the two loops interact in ways that destabilize the system. Once the inner loop is tuned and closed, the outer loop treats the inner closed-loop transfer function as part of its plant. This simplification is valid because the inner loop effectively makes its portion of the plant appear faster and less sensitive to variation.

A concrete example: in a shell-and-tube heat exchanger, the goal is to control the outlet temperature (primary variable) by adjusting steam flow. A disturbance might be a sudden change in steam supply pressure. In a single-loop arrangement, this pressure change alters steam flow, which slowly changes outlet temperature, and only then does the controller react. With cascade control, an inner loop measures steam flow directly and keeps it at the value commanded by the outer temperature controller. A pressure disturbance changes flow instantly, and the inner flow controller corrects it in seconds — before the temperature ever moves. The outer temperature loop simply commands what flow it needs, confident the inner loop will deliver it accurately.