Tuning of PID parameters for control valves

A few days ago I saw someone write something about Honeywell’s PID control; it’s quite classic. Here is a copy of the method for tuning PID parameters in Honeywell DCS control loops – the Honeywell PID parameter tuning method. 1. To modify PID parameters, one must have \"supv\" or higher permission levels. 2. Open the detailed screen of the control loop that is to be modified, navigate to the page shown in the diagram below, and adjust the three PID control parameters: K, T1, and T2 ； Meaning of PID parameters: K: Proportional gain (amplification factor), with a range of 0.0 to 240.0 ； T1: Integration time, ranging from 0.0 to 1440.0, in minutes; 0.0 indicates no integration ； T2: Differential time, ranging from 0.0 to 1440.0, in minutes; 0.0 indicates no differential action. IV. Functions of PID parameters (1) Characteristics of proportional control: 1. Fast response – as soon as a deviation occurs in the system, the controller amplifies this deviation by a factor of K and outputs it ； 2. There is a residual error in the system. The smaller K is, the smoother the transition process, but the larger the residual error ； As K increases, the residual error will decrease, but it cannot be completely eliminated; it can only serve a purpose of rough adjustment. However, if K is too large, the transition process is prone to oscillation, and excessive values of K can lead to divergent oscillations. (2) Characteristics of integral control: The output of integral control varies proportionally to the integral of the input error. Integral control can eliminate residual errors, but it reduces the stability of the system. As T1 decreases, the effect of integral control becomes stronger, and its ability to eliminate residual errors also increases; the output stops changing only once the error is eliminated. (3) Characteristics of differential control: The output of differential control is proportional to the rate of change of the process variable. By introducing differential action, the quality of control can be significantly improved. However, if the differential action is too strong, it can cause the control valve to switch between fully open and fully closed positions; therefore, T2 cannot be set too high. As T2 increases from low to high, the differential action becomes stronger, which has a significant effect on volume lag but no effect on pure lag. V. To understand how the control loop functions, one can access the detailed view of the control loop, going to the page shown below: Here, “CTLACTN” indicates the mode of operation of the controller, “REVERSE” denotes reverse action, and “DIRECT” denotes direct action. VI. Methods for selecting controllers (1) Selection of P controllers: It is suitable for systems where the lag in the control channel is small, the load changes little, and it is acceptable for the controlled variable to vary within a certain range ； (2) Selection of the PI controller: It is suitable for control systems with small lag, little variation in load, and where no tolerance for error in the controlled variable is allowed ； (3) Selection of the PID controller: It is suitable for control systems with large load variations, significant capacity lag, and high requirements for control quality, such as temperature control systems. VII. Methods for tuning PID parameters: In engineering applications, the trial-and-error method is generally used. The trial-and-error method is the most practical in practice. When setting the parameters, it is necessary to carefully observe the system’s response, and decide which parameters to adjust based on that response. To observe the system’s response, one can check the real-time trend curves in the detailed view of the control loop; it is ideal for the attenuation curve to have a ratio of 4:1, that is, the ratio of one peak to the next should be 4:1. Experience value: During actual debugging, it is only possible to set an approximate experience value first, and then make adjustments based on the resulting effects. Here, P represents the proportional gain, with P=1/K. Parameter range: Control system – P (1/K); K, T1/Min, T2/Min. Level: 20% to 80%; 1.25 to 5.0. Pressure: 30% to 70%; 1.43 to 3.4; 0.4 to 3. Flow rate: 40% to 100%; 1.0 to 2.5; 0.1 to 1. Temperature: 20% to 60%; 1.7 to 5; 3 to 10; 0.3 to 1. In general, during tuning, the system must not experience divergent oscillations; if such oscillations occur, it is necessary to switch to manual control immediately. Once the system stabilizes, reduce the gain, increase the integration time, or decrease the differentiation time, and then switch back to automatic control. The lower the magnification, the smoother the transition process, but the larger the residual error. The higher the magnification, the more likely oscillations occur during the transition process. The smaller the integration time, the faster the residual error is eliminated, but the system oscillations become more severe. The larger the integration time, the slower the system is at eliminating the residual error. If the differential time is too large, the number of oscillations of the system increases and the adjustment time lengthens; if the differential is too small, the system adjusts slowly. Steps for the trial and adjustment of controller parameters: Since proportional action is the basic form of control, it is necessary to first set the proportional gain appropriately. Once the transition process has stabilized to a certain extent, integral action is added to eliminate residual errors. Finally, derivative action is introduced to further improve the quality of control. The basic steps are as follows: (A) For a P controller, start with a low value for the amplification factor, then gradually increase K while observing the transition curve of the controlled variable, until the curve meets the desired standards ； (B) For the PI controller, set T1 to 0 first, and adjust the gain using pure proportional control so as to achieve a 4:1 attenuation curve ； Then K is reduced (by 10–20%), and the integration time T1 is increased step by step from large to small until a 4:1 decay profile is obtained ； (C) For the PID controller, set T2=0 ； First, use the PI control procedure to adjust the K and T1 parameters. Then increase the amplification factor by (10–20%) and reduce T1 accordingly; after that, gradually increase T2 from a low value to a higher one, while observing the transition curve until a satisfactory transition process is achieved. In short: when setting the parameters, it is necessary to carefully observe the changes in the system output and the variable being regulated, and then adjust the PID parameters accordingly based on the specific circumstances. It can be said that as long as process technicians spend a little more time, most control systems can meet the requirements using PID control. VIII. Tuning of the series pole control circuit: The tuning of the series pole control circuit can be carried out using a two-step method, that is, first tune the secondary circuit and then the primary circuit ； A one-step method can also be adopted, that is, tuning the main and auxiliary circuits simultaneously. (1) For tuning using the one-step method, the empirical values for the secondary loops are as follows; generally, proportional control is used only for these secondary loops: Secondary variable, Gain (K), Proportivity (P). For temperature: 5.0 to 1.7, 20% to 60%; for pressure: 3.0 to 1.4, 30% to 70%; for flow rate: 2.5 to 1.25, 40% to 80%; for liquid level: 5.0 to 1.25, 20% to 80%. (2) Bring the cascade control loop system into operation, and then adjust the parameters of the main controller using the parameter tuning methods applicable to single-loop control systems ； (3) If \"resonance\" occurs during the tuning process, it can be eliminated by reducing the amplification factors of the main and auxiliary controllers. If the resonance is severe, switch to manual mode first; once production stabilizes, restart the system and proceed with re-tuning. In summary: The P action is the most fundamental control mechanism; by adding the I action, error-free control can be achieved, improving control accuracy, while the addition of the D action enhances the overall quality of control. IX. Precautions (1) Before adjusting the parameters, it is necessary to verify the sensors and actuators to ensure that the instruments on site are functioning properly. Manual control can be used for testing; the parameters measured in manual mode should be stable and reliable. (2) Set the K parameter based on empirical values, temporarily turn off integral control and try to switch to automatic observation of the step response. At this point, special attention should be paid to the controller’s output; it is necessary to determine whether the circuit operates in negative feedback mode (check for any flaws in the design or wiring). This situation can occur during the debugging of new systems – for example, when it is necessary to adjust the opening degree, the regulator may instead increase that opening degree ； (3) When adjusting the parameters, it is necessary to ensure process stability; if the product quality or process parameters are affected, switch immediately to manual control, and only after the process becomes stable should automatic adjustment of the PID parameters be carried out. (4) In the case of cascade or proportional control loops, it is necessary to tune them one loop at a time, and attention should also be paid to the principle of tuning the inner loop first and then the outer loop. (5) When switching from manual to automatic mode, it is necessary to ensure that the set value is nearly identical to the measured value; however, this is not an issue with a Honeywell DCS system, as the set value in manual mode automatically tracks the measured value. X. Quick Reference for PID Parameter Tuning: To find the optimal parameters, check them in ascending order. Start with the proportional term, then the integral term, and finally add the derivative term. If the curve oscillates frequently, reduce the amplification factor. If the curve wanders around in large loops, increase the amplification factor. If the curve takes a long time to return to its normal state, decrease the integration time. If the waveform has a long cycle, extend the integration time further. If the oscillation frequency is high, first reduce the derivative term. If there is significant deviation and slow response, increase the derivative time. An ideal curve consists of two waves, with the first one being higher than the second one, in a 4:1 ratio. Observe carefully, make adjustments as needed, and conduct thorough analysis – this way, the tuning quality will be excellent. XI. Typical Control Circuits 1. Single-loop control: Principles for selecting the operating mode of control valves: Control valves can be classified into two types based on their operating mode: air-open and air-close. The air control valve increases its opening degree as the signal pressure rises; when there is no signal, the valve remains in a fully closed state ； Conversely, as the signal pressure increases, the valve gradually closes; when there is no signal, the valve remains in the fully open position, and this is what constitutes a air-operated shut-off valve. The selection principles are mainly: considering production safety, ensuring product quality, reducing losses of raw materials and energy, as well as the characteristics of the medium. (2) Controller operation mode selection: Selection principle: To make the entire single-loop system a negative feedback system. Specification: Control valve: air-open is “+”, air-close is “–”” ； Controller: Direct action is “+”, reverse action is “–”” ； Object: When the material or energy passing through the control valve increases, based on the process mechanism analysis, if the controlled quantity increases, it is denoted as “+”, and if it decreases, it is denoted as “–”” ； Transmitter: Generally considered a positive element. Then the discriminant for selecting the forward and reverse action of the controller is: (Controller “±”) × (Control valve “±”) × (Object “±”) = “–”. 2. Cascade loop control: As mentioned in single-loop control systems, the principle for selecting the forward or reverse action mode of the controller is to ensure that the entire control system functions as a negative feedback system; a discriminant based on a “negative product” is also provided. This discriminant is also applicable to the selection of the forward and reverse action modes for the primary and secondary controllers in a cascade control system. (1) Selection of the main controller’s operation mode: (Main controller ±) (Secondary object ±) (Main object ±) = (—). Therefore, when the main and secondary variables change in the same direction, the main controller should be set to reverse action mode; when they change in opposite directions, it should be set to direct action mode. (2) Selection of the secondary controller operation mode: (Secondary controller ±) (Control valve ±) (Secondary object ±) = (—) (3) Activation of the cascade control loop: By activation, it is meant to bring the primary and secondary controllers from manual operation mode to active operation mode through appropriate steps. Generally, there are two methods for putting a cascade control system into operation: one is to activate the secondary loop first and then the primary loop ； Another approach is to insert the main ring first and then the secondary ring. The first method is currently the one commonly used for commissioning. It is necessary to ensure a disturbance-free switch-over during commissioning; since the Honeywell DCS has a PV automatic tracking function, a disturbance-free switch-over can basically be achieved, and the commissioning process is relatively simple. 3. Split-range control system: A split-range control system is one in which the output signal of a controller is used to control two or more control valves, with each control valve operating only within a certain portion of the entire range of the output signal. That is, multiple valves with segmented strokes. There are mainly two methods for achieving this: one is through valve positioners mounted on each control valve, and the other is via DCS software. When DCS control is used, it is generally implemented through DCS groups in a proportional control system. Based on whether the control valves operate in an open-or-close manner in response to input signals, they can be divided into two categories: one category features valves that move in the same direction, meaning that as the input signal to the control valve increases or decreases, both valves open or close accordingly; the other category features valves that move in opposite directions, meaning that as the input signal increases or decreases, one valve closes while the other opens. The choice between co-rotating or counter-rotating operation of the control valve is determined entirely by the requirements of the process. As shown in the figure below: The ratio control loop is a control system that automatically maintains a certain ratio between two or more quantities of materials, so that these quantities change in proportion to the main material quantity. For example, in a waste heat boiler, the amount of gas and the amount of air supplied are in proportion to each other, thereby ensuring optimal combustion performance of the boiler. (1) Types of ratio control systems:

http://bbs.hcbbs.com/viewthread.php?tid=560113&page=1#pid3154016