PID dynamic parameter change
Thread Content
Dear marine engineers, I’m sure everyone has used PID control in practical projects. Have you ever noticed that when using PID control, the selection of the P, I, and D parameter values is very important for a specific control object? These three parameters directly determine the operating state of that control object. So, I would like to ask whether the P, I, D parameters in PID control can be changed dynamically. Note that this is a dynamic change: during the PID control process, it is possible to adjust one of the P, I, or D parameter values as needed, while the control process continues. Friends who have played it, please share your thoughts.
The adjustment of PID parameters involves continuous optimization during the actual control process, with the goal of finding the optimal values for these control parameters.
Textbooks describe a standard parameter tuning process, step by step; if the requirements are not high, it is also possible to use empirical values for tuning.
You mean online editing, right? All DCS systems support this feature.
Previously, PID parameter tuning was done based on human experience, but now smart meters are used, which have a self-learning function, making the process much simpler.
The poster is referring to the real-time changes in the P, I, D values
Frequent changes can confuse people. Traditionally, PID control is linear. In the entire control loop, aside from the fact that the process itself may be nonlinear, the transmitters and positioners along with their travel ranges are linear; therefore, equal percentage valves are commonly used to \"correct\" the nonlinearity of the process. According to control theory, it is a necessary condition for the stability of a control loop that the gain product of the entire loop be around 0.5; therefore, we generally set a set of PID parameters for the condition of 100% load. However, when the load changes, the gain of the process also changes; if it deviates from the baseline of 0.5, the control accuracy decreases. If the process is considered acceptable, there is no need to modify the PID parameters. PID is used in conjunction with the characteristics of a process; for level control, one PID controller is sufficient, as the process is linear and the control requirements are not high. If two sets of PID parameters are to be prepared, they must be intended for two separate process streams; for example, one controller is used for both heating and cooling ； Or, operations with frequently changing loads.
Emerson has a system for tuning PID parameters, but it is based on an analog process: the simulated values of these parameters are compared with the actual values during operation, and reasonable parameter settings are provided to the customer for reference. Dynamic adjustment is generally not carried out, especially for those parameters that are particularly sensitive.
This post was last edited by jlshnlhj on 2019-10-21 at 10:02; it is possible to \"dynamically change\" the P, I, D values starting with pneumatic instruments.
What he meant was not PID tuning, but tracking
With follow-up control, we have encountered similar operating conditions as well. It is no longer fixed-value control in the traditional sense; there are discussions about it online, but I only have a vague understanding of it – it’s known as fuzzy control. Some expert can give some extra tutoring.