The automatic regulator setting was implemented, although the measurement was not stable. If the measured change is large relative to the reaction of the actuating pulse, then the test results will be distorted.

1 

2 

The reaction will not be stabilized before returning to the original manipulated variable. The calculated parameters are therefore false.

The reaction of the process to the actuating pulse is insufficient relative to the level of noise/interference. The measurement should be filtered or step_ampl should be increased.

tmax specifies the frequency with which the measurement is taken, i.e. the value that is used to calculate the coefficients. tmax must be between 1 and 5 times the rise time of the repeated task.

This bit is used when the reaction to an actuating pulse significantly exceeds (overshoots) the measured value (i.e. by more than 10%). The process does not conform to the models used by the algorithms.

This bit is used when the reaction to an actuating pulse leads to inversion of the reaction at the initial stage (i.e. undershoots by more than 10%). The process does not conform to the models used by the algorithms.

The reaction of the process is asymmetrical.

The last parameter set must be a compromise between the reactions at ascent and descent. Both cases concern average performance.

If the desired criterion is the length of the reaction on ascent, then the first parameter set must be taken into consideration. During the return phase (to the original manipulated variable) the automatic regulator setting is turned off. If the desired criteria is the length of descent, then a negative amplitude must be used.

The process includes an integral component or tmax is too small and the process asymmetrical. The calculated coefficients must correlate to the process with the integral coefficient If this is not the case, the automatic regulator setting should be restarted, after tmax has been increased.