Why is PID not works good under noisy sensor feedback condition, while it performs very well if system is externally disturbed?
Actually PID control system is try to reduce error signal which is the different between reference signal and feedback signal, when noise (which is an always random and uneven signal) is added in PID based F.B loop control system, then the error signal will also become uneven (alter in every instant of time) so fix PID gains would not be able to settle it to minimize it.
While external disturbance is normally act as some sort of extended impulse and there is no unevenness in it , so PID can easily overcome it and makes system stable.
If you want to remove noise by using PID controller, you should add some sort of filter (i.e: Kalman Filter or Extended Kalman Filter) on feedback response of system.
PID controller isn't a good filter, because as Syed Faize Ahmed wrote its job is to minimize a difference between SetPoint and a current Process Value. You can easily check it using any simulation tool e.g. simulink and mix a process value with e.g. a sinus signal. Generally external disturbances usually called loads, which are related to the technological assumptions e.g. in the level control system a control value i.e. a stream of liquid from the pump is loaded by a stream which comes outside. Such disturbances i.e. loads are often measurable. Unfortunately disturbances related to sensors usually called noises, are stochastic.
Theoretically we can use a model of the noise during the controller synthesis and try to make controller able to remove its influence on the control value, but in practice a typical industrial control device e.g. a PLC controller doesn't have an opportunity of implementation of the inverted model in the base controller algorithm.
The only solution is as Muhammad Tanveer wrote is using filters and usually analogous filters. In my opinion a control engineer should select only such measurement devices which enables filtration on their board. I think that a fight against noises in the digital machine is very hard usually because of the fact of sampling effects which can lead to change high frequency noise into the low frequency noise.
The question is : why PID needs D mode ? The solution with no D mode is possible to decrease an influence of noise. Unfortunately sometimes we need D mode : in case of a narrow stability margin with P or PI controllers. Practically in such situations even a little value of proportional gain can make a control system unstable. Such system are called systems with a narrow band control.
I have one small question:
1)As Muhammad sir said the only way to remove noise is adding a filter in the feedback.
But noise itself don't have definite frequency I mean It is random.How will we design the filter so that it remove the noise in the feedback?
2)How will we model the noise when we are synthesizing the PID controller?
To my understanding, if the noise is external to the system, it can also be considered as a disturbance. However, this does not change the linearity of the system. Whereas, if a sensor in the feedback look is noisy, it can affect the transform the signal to a nonlinear domain. Thus, probably this reduces the efficiency of the controller. And yes differentiator, amplifies the noise. However, I am skeptical, if a system with a noisy sensor and no differentiator will also produce a desired rejection of noise.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps