Hello guys,
I'm pretty much bored this weekend and since there is snow outside, I decided to start new thread about PID's control design practice. Recently, I had an opportunity to speak with one of the Siemens' control design experts. It was a very brief conversation, since his time is very expensive. However, I started thinking about different practices when it comes to design of simple control loops.
It is interesting that this guy acts only as negative gains don't exist. What I mean by this is that controller is always design with all positives gains (or gain and time constants depending on controller's form), and depending of the nature of a process, error signal is defined as SP-PV or PV-SP.
Basically there are two types of control loop (example of controlling level in some kind of tank):
1. "Inlet closed loop" – this is the case when level is controled with control valve at input of a tank. In such case, error signal is defined as SP-PV (setpoint – process value), because when SP is grater then PV, error is positive, and since gain is also positive, valve will open more, allowing more fluid to flow int tank.
2. "Outlet closed loop" – this is the case when level is controled with control valve at the bottom of a tank. In that case, error signal is formed as PV-SP. In this case when error is >0, since gain is >0, valve must open more allowing more fluid to flow out of the tank.
Now, as you probably noticed some kind of convention is used according which, always, when error is positive, actuator must go opening. This guy said that when things are defined in such way (convention) there is very little room for confusing especailly when it comes to control loops designs in power plants.
I'm interested to learn what is your practice.
When I have time, I'm analysing implementation of some simpler control loops at work. Usually, it's about level control. On few places I have found very interesting control loop implementation.
Input to the PID controller is formed this way: (SP-PV)*Control_Valve position*Constant. Basically, error signal is multiplied with control vave_Position (0-100%)and with constant (typically in range 0.01 - 0.1).
That is very strange to me, and I'm not sure what is the possible purpose. Did you see something similar?
I'm pretty much bored this weekend and since there is snow outside, I decided to start new thread about PID's control design practice. Recently, I had an opportunity to speak with one of the Siemens' control design experts. It was a very brief conversation, since his time is very expensive. However, I started thinking about different practices when it comes to design of simple control loops.
It is interesting that this guy acts only as negative gains don't exist. What I mean by this is that controller is always design with all positives gains (or gain and time constants depending on controller's form), and depending of the nature of a process, error signal is defined as SP-PV or PV-SP.
Basically there are two types of control loop (example of controlling level in some kind of tank):
1. "Inlet closed loop" – this is the case when level is controled with control valve at input of a tank. In such case, error signal is defined as SP-PV (setpoint – process value), because when SP is grater then PV, error is positive, and since gain is also positive, valve will open more, allowing more fluid to flow int tank.
2. "Outlet closed loop" – this is the case when level is controled with control valve at the bottom of a tank. In that case, error signal is formed as PV-SP. In this case when error is >0, since gain is >0, valve must open more allowing more fluid to flow out of the tank.
Now, as you probably noticed some kind of convention is used according which, always, when error is positive, actuator must go opening. This guy said that when things are defined in such way (convention) there is very little room for confusing especailly when it comes to control loops designs in power plants.
I'm interested to learn what is your practice.
When I have time, I'm analysing implementation of some simpler control loops at work. Usually, it's about level control. On few places I have found very interesting control loop implementation.
Input to the PID controller is formed this way: (SP-PV)*Control_Valve position*Constant. Basically, error signal is multiplied with control vave_Position (0-100%)and with constant (typically in range 0.01 - 0.1).
That is very strange to me, and I'm not sure what is the possible purpose. Did you see something similar?
