PID tuning

This question is too general to answer. Whether a self-regulating loop is stable depends on various factors: 1. Whether the process itself can be easily stabilized and controlled. 2. Whether the process control strategy is appropriate. 3. Check whether the detection point is appropriate. 4. Check whether the measuring instrument is operating stably. 5. Whether the selection of the control valve is appropriate and whether the positioner is stable. 6. Are the PID parameters appropriate? ...

First, remove the derivative term and observe the curve; if it remains the same, then change the scale (increase or decrease it) and examine the valve opening curve. If there is an improvement, continue to increase the magnitude of the scale changes. Once a significant improvement is seen, then adjust the integral term (increase or decrease it), and observe the curve until control becomes stable. Adjusting PID parameters is a process of gradual tuning; it may take a whole day to get the parameters of a relatively complex control system adjusted properly. Additionally, when tuning the PID parameters, the system must be operating stably. Personal opinion, for reference only

The larger the P value, the more sensitive the curve of the control valve becomes. The graph above seems to indicate that the P value is too high; it should be reduced

Let’s check the process requirements. It depends on the fluctuations in the actual values.

This post was last edited by jlshnlhj on 2020-7-2 at 12:07. 1. If PV is not temperature, then there is no pure lag in heat transfer, and overshoot is not necessary; in such a case, adding differentiation only serves to cause confusion. 2. If the operator wants such a curve, there’s nothing to discuss. 3. In the graph, the PV and MV curves oscillate within a small range with roughly constant amplitude, indicating that either P or I is at a critical value; the amplitude of MV is greater than that of PV, and the value of the adjustment parameter K is not large. Due to unclear target of adjustment (pressure? Liquid level? Vertical tank? Horizontal tank? ), the time unit is also unclear, so it’s hard to determine what caused it. Based on the analysis of the PV and MV curves combined with the P, I, D parameter values, it can only be estimated as I. 4. The unit of P is %, which represents the degree of proportionality, not the proportional action. Unit I is S, and D is also S. 5. If you want to make some adjustments, try: P:75-120%, I:90-180S, D:0S. Then look at the curve and make adjustments.

Common mnemonic for PID control: To find the optimal parameter settings, check values in ascending order. Start with the proportional term, then add the integral term, and finally the derivative term. If the curve oscillates frequently, increase the proportional gain; if the curve wanders around in large loops, decrease the proportional gain. If the curve takes a long time to return to its normal position, reduce the integral time. If the fluctuations in the curve have a long cycle, extend the integral time further. If the oscillation frequency of the curve is high, first reduce the derivative term. Large errors lead to slow fluctuations. The differential time should be increased. For the ideal curve, there are two waves – higher at the beginning and lower at the end, in a 4-to-1 ratio. Observe carefully, make adjustments as needed, and conduct thorough analysis; this way, the quality of the adjustments will not be poor. Think about it, consider it carefully.