D
I had hoped that someone would mention the alternate form of bumpless transfer, but I guess not. So I will.
In cascade control, the setpoint of a secondary control loop is determined by the output of the primary controller. Sometimes, the
secondary loop may be placed into manual so that the operator can take over manual valve positioning. When the secondary is returned
to CAS, we want that change to be bumpless. This is accomplished in all control systems in two ways:
1) allowing the setpoint to track the PV in manual, or
2) simply removing the reset remainder, derivative term, and proportional term, allowing the upstream setpoint to remain.
The reason for method 2) is to allow the process to come to equilibrium with the setpoint calculated from an upstream cascade
without the overcontrol that would result from using the wound-up reset and the proportional kick when AUTO mode is resumed. This
allows the control to resume at a more gradual rate without the sudden movement (bump) that would be caused if the proportional term
and derivative term were to be used on that initial computation. We speak of this as allowing the control to resume under control of
the reset term. All commercial PID controllers offer these two options. Both methods are said to be "bumpless."
Dick Caro
In cascade control, the setpoint of a secondary control loop is determined by the output of the primary controller. Sometimes, the
secondary loop may be placed into manual so that the operator can take over manual valve positioning. When the secondary is returned
to CAS, we want that change to be bumpless. This is accomplished in all control systems in two ways:
1) allowing the setpoint to track the PV in manual, or
2) simply removing the reset remainder, derivative term, and proportional term, allowing the upstream setpoint to remain.
The reason for method 2) is to allow the process to come to equilibrium with the setpoint calculated from an upstream cascade
without the overcontrol that would result from using the wound-up reset and the proportional kick when AUTO mode is resumed. This
allows the control to resume at a more gradual rate without the sudden movement (bump) that would be caused if the proportional term
and derivative term were to be used on that initial computation. We speak of this as allowing the control to resume under control of
the reset term. All commercial PID controllers offer these two options. Both methods are said to be "bumpless."
Dick Caro
