How are the PID parameters of a control valve determined?
Thread Content
To achieve relatively accurate automatic control in a control loop, appropriate PID parameters are required. So, what factors determine these parameters? How is it determined?
When configuring DCS, there should be a default value first, and for most control valves, no modifications are necessary. All that’s left is to make adjustments as needed.
The adjustment of PID parameters should be done gradually on-site by increasing or decreasing them; there is no universal standard or formula for this. The optimal PID values vary from one system to another. When making adjustments, an approximate empirical value can be used, followed by parameter tuning based on the response curve. If the output fluctuations are large, the effect of the proportional and integral terms can be reduced appropriately; if the fluctuations are slow, these terms can be increased. In general, the derivative term is not included during parameter tuning to avoid causing oscillations in the system.
The PID parameters of the control loop need to be tuned, but in DCS configuration, empirical values are used initially, which are then adjusted based on the quality of control.
In other words, it’s unknown how it came about. Many of our control valves cannot be automated; for example, when trying to maintain a temperature of 70 degrees, the value may fluctuate between 50 and 90. There is no other option, so in the end, manual adjustment is the only choice.
I see everyone has mentioned configuration, but I’m not quite clear about what the concept of configuration is. Some control valves are installed later; on the DCS screen, their information is copied from previous control valves, and it’s not clear whether a so-called configuration process was carried out.
Thank you so much! Proportional and integral control: which one controls the amplitude of the fluctuations, and which one controls their speed? Or is there some connection with both?
Are the PID parameters different for each loop? How are the experience values for each circuit determined? If the default experience value control quality is not satisfactory, experimental adjustments will be required, which seems like a considerable amount of work.
It is possible to control the amplitude and speed of fluctuations; the difference lies in the mathematical models used. The proportion is adjusted based on the magnitude of the deviation – the greater the deviation, the larger the output. When the deviation is eliminated or reduces to a certain level, the proportional term produces no output. Integration operates depending on whether there is a deviation or not; as long as a deviation exists, the integral term will generate an output, and the magnitude of this output depends on the integration coefficient and the size of the deviation. Differentiation adjusts according to the rate of change in deviation; the faster the deviation changes, the greater the output, and no output is produced when there is no change, thus serving as a form of lead regulation.
High requirements are placed on regulating quality, and the PID parameters for each circuit generally vary. Experience value refers to the parameter values that are determined based on a lot of trial and error, derived from intuition. If you’re not satisfied, just make adjustments.........
To be honest, although PID consists of just three simple parameters, it’s truly amazing if one can adjust them properly. There are people who specialize in this; how much do they charge per circuit?