pHman

Controller for the pH Control ApparatusWritten by Alma Lincoln

Spring 2007

In the pH control experiment of the chemical engineering senior unit ops laboratory, streams of aqueous nitric acid (HNO3) and sodium hydroxide (NaOH) are combined in a continuous stirred tank reactor in order to neutralize the pH of the resulting mixture. The valves regulating flows are controlled electronically via a LabVIEW interface, which is equipped with PID controllers to regulate reactor pH and liquid level with respect to flows, as well as to control flows with respect to valve settings. The purpose of this experiment is to familiarize the student with the process of automated PID control on a system exhibiting nonlinear dynamics. This manual is intended to instruct students on the use of the LabVIEW interface for the experiment.

LabVIEW User Interface

The following paragraph is taken from the document “Operation Manual” located in the pH control apparatus computer, which includes a detailed description of the apparatus as well as in-depth instructions.

“The LabVIEW user interface allows the operator to manually, or automatically control three inlet valves and one outlet valve, while recording pH, flow, and vessel level. The interface is divided into five sections responsible for recording and sending the required data. The five sections are flow diagram, relative time plots, valves, PID controls, and data logging.”

A diagram of the interface is shown on the following page. Sections of the diagram are numbered in red, and the functions of these sections are described subsequently. The interface is accessed from the desktop of the apparatus computer. There it appears as an icon labeled “Shortcut to pH Cascade ’04.”

1. pH and level controllers – These sections contain the PID controllers for the reactor pH and liquid level. The user must select which variables will be used to control pH and level. Level may be controlled by acid, base, or outlet flow, while pH may be controlled by acid or base flow. The same flow cannot be used to control both variables. Once the variables are selected, the user specifies setpoints and the controller parameters K, τI, and τD. The user can also activate data filtering to reduce measurement noise, and activate or deactivate cascade control. This will be explained in the next section.

2. Flow controllers – These sections contain PID controllers for the flow rates of the acid, base, buffer, and outlet streams. These controllers are used during cascade control: a setpoint is specified for the reactor pH or level, and the controller maintains this setpoint by determining a necessary flow rate for a particular stream. At the same time, another controller regulates the flow of that stream by varying the valve setting, using the flow rate output of the first controller as a setpoint. This counteracts the effects of valve sticking and hysteresis. In these blocks the user may only specify setpoints and control parameters; the choice of variables is fixed.

3. Valves – This block displays the valve settings. These have units of percent open, ranging from 0% to 100%. These settings may be set manually by the user when control is inactive, or when a particular valve is not used by a controller. When a valve is involved in control, its setting is displayed here.

1 2

3

4

5

6

7

8

4. Relative time plots – Data are plotted in this section. The top plot displays the setpoint and measured value for the pH over time. Below is a similar plot for reactor liquid level. The user may adjust the axis scales of both plots.

5. Data logging – In this section the user records and saves data. The user must specify a name for the file and a time interval for recording data. The default setting is one data point every ten seconds, though the user will likely prefer to collect data more often. Data may be collected as often as one point per second. Files are saved in the directory C:\pH data\ as .txt files, and may be imported to Microsoft Excel for analysis.

6. Shut down – This button is used to shut down the interface in order to exit the program. The button needs to be held down for several seconds in order to end the program. Manual valves need to be closed before shutting down.

7. Flow diagram – In this block, a flow diagram of the system is displayed. It shows schematically how the acid, base, and buffer tanks are connected to the reactor, and how the reactor is connected to the waste tank. The flows of the streams connecting all the vessels are displayed here, along with the liquid levels and pHs of the reactor and waste tank. Inside the picture of the reactor are buttons used to plot the reactor pH and level.

The pH of the waste tank must remain within a certain range, between 5 and 9, for safe disposal. If the waste tank pH leaves this range, the waste pH indicator blinks red as a warning. In this event, the contents of the tank must be neutralized before disposal.

8. Waste tank warning – This area of the interface is where the waste tank warning appears. This happens when the waste tank becomes too full, around 28 gallons. When this occurs, all flows are shut down until the tank is drained. The drain is located on the wooden platform, directly in front of the tank.

LabVIEW User Interface with Frequency Response

In addition to the interface described previously, there is an alternative version of the interface which may be used to perform a frequency response analysis. In such an experiment, the setpoint of a controlled variable is not held constant. Instead it is varied as a sinusoidal function of time. This interface is also located on the desktop, where it appears as an icon labeled “Shortcut to pH Frequency Level 2.” This is not to be confused with “Shortcut to pH Frequency Level,” which is an earlier version of the interface that only allows frequency response analysis on the level controller and not the pH controller. It is displayed below.

This interface has a layout similar to the one previously described. However, it includes an additional section, outlined here in red. In this section the user selects which setpoint to vary. The options are pH, tank level, or “NONE” (meaning frequency response is inactive.) Thus frequency response analysis may only be performed on one variable at a time. When frequency response is active, the setpoint value specified in Block 1 is used as the “zero” value around which the setpoint of this variable will oscillate. Below the variable field, the user specifies the frequency of oscillations. Below the frequency the user specifies the amplitude of oscillation, which has the same units as the controlled variable. At the bottom of this section, the current value of the setpoint is displayed.

PID Control

This interface includes six PID controllers, all of which implement the positional form of control:

pk=p¿+Kc [ek+

Δtτ I

∑j=1

k

e j+τ D

Δt (ek−ek−1) ]An in-depth discussion of this equation can be found on page 201 of Process Dynamics and Control, Second Edition, by Seborg, Edgar, and Mellichamp. Note that this equation is based upon discretely sampled values taken at intervals of Δt; it is not a continuous function. Thus, pk is the controller output at time interval k. The proportional term, ek, is simply the error, or difference between setpoint and measured value, at interval k; the

other two terms. The integral term is defined as the cumulative sum of errors up to interval k, multiplied by the length of the interval and divided by the appropriate time constant. The derivative term is defined as the difference between the error at interval k and that at the previous interval, dividing by the length of the interval, and multiplying by the appropriate time constant. The PID equation must be defined this way due to the noncontinuous nature of data measurement.

In the block diagram for this interface is a series of frames, each of which handles a different task. These tasks include data collection, control, plotting, saving of data, and other functions. This manual will focus mainly on frames relevant to PID control. The first two frames handle importing data from the FieldPoint software. The first frame, frame 0, collects pH data:

The visual interfaces (VIs) that communicate with FieldPoint are seen on the left side of the diagram, outside the loop. Here, the one on the top sends pH data into the loop, as shown in this detail:

Note the feature highlighted by the red box: the data streams connect to the loop, meaning that the data are carried through each loop of the program. This indicates that the positional form of control is in use, as the looping of data allows the continuous summation of error in the integral action.

Frame 1 shows the acquisition of level data:

Again, a detail:

Frames 6 and 7 handle control: 6 controls the reactor pH and liquid level, while 7 controls the flow rates. Diagrams of both frames appear below, 6 before 7.

The PID controllers, which appear as blue boxes, are separate VIs. These VIs calculated the required output based on the positional equation discussed above, using the control parameters entered into the interface by the user. The outputs from frame 6 are sent as inputs to frame 7 for cascade control in the lower right of the frame. In frame 8, controller outputs are defined as valve settings. Outputs from frame 6 are used for regular control (seen in the lower right corner), outputs from frame 7 for cascade control (upper right corner). On the following page is a detail of frame 8, in which the acid and base valve settings can be seen.

The valve settings are then sent to frame 9:

Here, they are sent to a VI, labeled “pH Valves,” that converts valve settings to electronic signals. This VI also takes in the current valve settings from a second FieldPoint VI, then sends the new settings back to FieldPoint. A detail of the valve VI is shown on the next page.

The block diagram for the frequency response interface is similar to that of the original interface. However, two major differences exist. First, an additional frame is added to the frequency response diagram. This frame defines a sinusoidal function to be used as a setpoint, and lists what variables may use this function: here, reactor pH and level. The function must take in two inputs, frequency and amplitude, which are supplied by the user. A detail of the sinusoidal function, which is in frame 6 of the interface, appears below.

Due to the addition of this frame, the functions of subsequent frames are moved back. For instance, frame 7 of the frequency response interface does what frame 6 of the original interface did.

Secondly, in order for the interface to make use of the sinusoidal setpoint, all instances of the local variables “pH Set Point” and “Level Set Point” are set to take in the function. These variables occur in three frames, shown below for comparison: frame 2 (also frame 2 in the original interface),

vs.

frame 7 (frame 6 in the original interface),

vs.

and frame 10 (frame 9 in the original interface).

vs.

Reference

Seborg, D., Edgar, T., Mellichamp, D. Process Dynamics and Control. 2nd Edition. Wiley Publications. 2004.

Operation Manual.