Figure 3Industrial batch reactor P&I diagram
Most of old-fashion PID tuning methods require PID data from which clear step change of PID OP in manual mode is visible or data showing oscillatory PID controller behavior, as shown in Figure 4. All these tests are time consuming, especially in the case of a slow TC, and they can seriously harm the process and produce off-spec product. Based on the conducted step-tests in manual mode the user will often need to calculate (visually ) several parameters from the trends, depending on the used old-fashion PID tuning method.
Trial-and-Error method will require to know time-to-steady-state (TTSS) based on which the process time constant (parameter which explains the speed of the response ) can be calculated. This parameter can be correlated with integral tuning parameter, but P&D parameters will be set just as a guess value based on the user past experience, as shown by Table 1. For using closed-loop Ziegler Nichols PID tuning method user needs to mark ultimate proportional gain and ultimate oscillation period. For open-loop Ziegler Nichols PID tuning method the user needs to calculate process reaction curve and its parameters. Relay method needs high and low peak values of PID OP which made PID controller to oscillate in manual mode. Lambda method requires to know closed-loop dead time and closed-loop time constant. IMC method needs to have open-loop dead-time, gain and time constant, based on visual trend analysis. Some other methods require the calculation of decay ratio (typically set as 1:4 ).