SaskDreamTeam
Member
Hi Everyone,
I am working on a project in Proworx NXT, and noticed something with some in service PID controllers. I have a Flow Controller configured in a Low Selector with a Discharge Pressure Controller on a pump. Some logic is written that calculates the PID error (externally from the PID function), and this error is used to determine which PID has the lowest error, and selects that PID for use.
In the logic, a 20% of full PV range (for example, 0 to 500 m3/hr flow signal would = 100 m3 bias) is calculated. This PV 20% Bias is subtracted from the setpoint before the SP is used in an error calculation.
From what I can tell, this bias would increase the calculated errors respectively, amplifying them. This error calc does not affect the Controller Output as it is external, and at no point in time is a bias constant added to the controller output (which is my understanding of how Bias is usually used).
I am wondering if anyone has ever seen this type of logic or control before. Is there any fundamental value in Biasing an error calc for a low selector? I appreciate anyone's insight or suggestions. Thanks in advance. :site:
I am working on a project in Proworx NXT, and noticed something with some in service PID controllers. I have a Flow Controller configured in a Low Selector with a Discharge Pressure Controller on a pump. Some logic is written that calculates the PID error (externally from the PID function), and this error is used to determine which PID has the lowest error, and selects that PID for use.
In the logic, a 20% of full PV range (for example, 0 to 500 m3/hr flow signal would = 100 m3 bias) is calculated. This PV 20% Bias is subtracted from the setpoint before the SP is used in an error calculation.
From what I can tell, this bias would increase the calculated errors respectively, amplifying them. This error calc does not affect the Controller Output as it is external, and at no point in time is a bias constant added to the controller output (which is my understanding of how Bias is usually used).
I am wondering if anyone has ever seen this type of logic or control before. Is there any fundamental value in Biasing an error calc for a low selector? I appreciate anyone's insight or suggestions. Thanks in advance. :site:
