With the increase in processor power, the number of modern Advanced Process Control (APC) algorithms available to users has reached a level where everyone should apply them if there is an opportunity to improve their operations. However, in some older plants, due to still low processor power usage and high business financial constraints a cascade or feed forward control loops could be considered also one of the APC strategies. So what really constitutes this APC concept? Let’s look at its three basic elements.
Dead time compensation is one of the first Advanced Process Control (APC) concepts that are still not very well understood. Dead time in the process is mostly caused by transportation delay, like hold-up in trays, down comers, distributors, piping and so on. In the literature, it can be easily found as Delay time, Pure time, Time delay, Transport delay, Distance-velocity lag and so on. Many think that simply inserting a dead time function block into the input to the loop is adequate, but doing so neglects the effect of process disturbances to the output of the PID controller. Utilizing the most famous dead time compensation function, called Smith Predictor, allows the control loop to adjust the model bias according to the magnitude of the disturbance. A modified Smith Predictor also allows the controller to adjust the gain as well as the bias depending on the disturbance. The model in the Smith Predictor can be identified Transfer Function model, any 1st principle model, and any identified empirical model with multi-inputs or any other capable model for predicting the main controlled process variable.
A second Advanced Process Control (APC) concept that is implemented using standard function blocks is the adaptive tuning, sometimes called gain scheduling. Any PID or APC control loop that exhibits changes in the process response that are essentially linear for that portion of the operating range is a candidate for this specific concept. Actually, changes in process or operating conditions which causes the transfer function model parameters to change a lot (more than 20%) can be a good candidate for mentioned APC concept. Transfer function model parameters can usually change because of: change in throughput, changes in operating conditions and changes in control valve position (<10%, >90%) which will cause cycling (aggressive or fast control) or sluggish control (slow control).
The implementation of this isn’t difficult but the user does need to know how many control regions there are and where the changes occur in the response of the PID or APC concept control loop. The user will need to use a PID or Advanced Process Control (APC) concept tuning software package to determine the responses in each of the linear regions.
The third Advanced Process Control (APC) concept recognizes control loops and/or process signals which interact or fight to each other, so the design of a feedforward concept is one of the useful APC algorithms. It requires measurements of disturbance and controlled variables in the process. The complete procedure consists of two steps: first to calculate the effect of disturbance on the controlled variable and then to determine the required movement of the manipulated variable to cancel measured disturbance. Conceptually, the disturbance signal is taken through dead time, feedforward gain and lead/lags function blocks in series and summed with the output of the PID control loop. Benefits of using this APC approach are huge. The benefits of feedforward control compared to feedback control lie in the proactive action by removing the disturbance effect completely at the beginning of its entrance in the process and creating stable process conditions.
PITOPS software from PiControl Solutions LLC identifies multivariable transfer functions from process data. Once PITOPS identifies the two transfer functions – one from MV to CV and the other from DV to CV, then PITOPS can mathematically calculate the feedforward transfer function parameters – lead, lag, gain and delay. Estimation of these four feedforward tuning parameters is typically done by guesswork with little chance of being correct. Guessed and wrong feedforward parameters will produce out-of-phase control action that will be harmful and better than no feedforward control scheme at all! PITOPS methodology is simple, practical and helps to design, implement and tune feedforward control schemes fast, easily and most importantly – correctly!