Question 13 CEL01 - Chief Engineer - Limited

The Correct Answer is C **Explanation for Option C (Correct):** The scenario describes a "Proportional Only" (P-only) controlled loop that is "oscillating continually, above and below the setpoint." In control systems, continuous oscillation indicates that the controller's proportional gain (P-gain) is set too high—it is likely at or near the ultimate gain ($K_u$). To stabilize an oscillating loop, the gain must be reduced. Decreasing the gain makes the controller less aggressive, damping the oscillations. The "quarter wave damping" response (where each successive peak is one-fourth the amplitude of the preceding peak) is the generally accepted optimal tuning criterion for stability and speed in industrial PID (or P-only) loops, as formalized by methods like Ziegler-Nichols. Therefore, gradually decreasing the gain is the necessary action, and stabilizing the loop into a quarter-wave response is the desired outcome. **Explanation of Why Other Options Are Incorrect:** * **A) By increasing gain, the system's oscillations should subside vs. setpoint after an upset.** * This is incorrect. Increasing the proportional gain makes the controller more aggressive. Since the loop is already oscillating (indicating the gain is too high), increasing it further will only destabilize the loop more, making the oscillations larger or turning them into runaway instability. * **B) By decreasing gain, the process should return to a straight-line response vs. setpoint after an upset.** * This is incorrect regarding the expected outcome. While decreasing the gain is the correct action, a straight-line response (perfect critical damping, $K_c = K_u/2$) is difficult to achieve consistently and generally results in a slower response than desired. The typical goal for optimal stability and speed is the quarter wave response, not a perfectly flat, critically damped response. Furthermore, a P-only controller will always settle with an offset (proportional droop) relative to the setpoint when there is a change in load, so it cannot return to a "straight-line response vs. setpoint" unless the load returns to its initial condition. * **D) By decreasing reset, the system's oscillations should subside vs. setpoint after an upset.** * This is incorrect because the problem specifies a "Proportional Only" (P-only) loop. P-only controllers do not have a reset (integral) term to adjust. Even if it were a full PID controller, "reset" (integral time, $T_i$) is typically measured in time (e.g., minutes per repeat or seconds), and decreasing the numerical value of $T_i$ actually *increases* the integral action (making the controller more aggressive), which would worsen existing oscillations.