Under what circumstances are the proportional, integral, and derivative controls of the regulator adjusted?

If there is a deviation, the integration time should be reduced, as integration can eliminate residual errors. If the system has lag, differentiation can be introduced appropriately by increasing the differentiation time, since differentiation has a leading effect

I. There is a mnemonic: To determine the optimal parameter settings, check in ascending order; start with proportionality, then integration, and finally add differentiation. If the curve oscillates frequently, increase the proportionality setting; if the curve wanders around in large loops, decrease this setting. If the curve takes time to return to its normal position, reduce the integration time. If the fluctuations in the curve have long periods, extend the integration time further. If the oscillation frequency of the curve is high, first reduce the differentiation value. If there is significant error and slow response, increase the differentiation time. An ideal curve consists of two waves, with the first one being higher than the second one, in a 4:1 ratio. Observe carefully, make adjustments as needed, and conduct thorough analysis – this way, the quality of the adjustments will be excellent. II. There is an empirical data adjustment method that is often used. Temperature T: P=20~60%, T=180~600s, D=3-180s. Pressure P: P=30~70%, T=24~180s. Liquid level L: P=20~80%, T=60~300s. Flow rate L: P=40~100%, T=6~60s.

Based on my engineering experience gained while working in SIEMENS’ engineering department, these parameters can be calculated directly from the parameters of the control object. Nearly 10 years of practical experience have fully demonstrated their effectiveness. This is why the controllers used by those large companies are able to optimize their parameters automatically; such methods have been standardized and turned into specialized algorithm libraries, but they are not sold as part of standard algorithm libraries!