Liquid Level Control of Gravity Drained Tank

Liquid Level Control of Gravity Drained Tank

Lab Procedure

Goal: To investigate the dynamic behavior of gravity drained tanks and design control strategies to maintain the level through feedback systems.

In this experiment, you will study and design the control system of a liquid level in a 10 liter tank in the presence of feed flow rate and outlet flow rate disturbances. The self-regulating behavior of the process means that the tank will seek a steady operating level if the inlet flow and disturbances are held constant.

The cart is designed to use water from the lab faucet and is controlled by CV001, which then flows through a MicroMotion Flow Tube into the 10 liter vessel. The rate of water draining from the tank through the 0.25” ID hose is controlled by CV002. The water level in the tank is measured by a Milone eTape level transmitter that acts as a variable resistor providing instantaneous measurements to the microcontroller. Only the Chemical Engineering Lab Technician is permitted to touch or remove the electrical tape from any of the devices on the apparatus.

Day 1

Before starting the experiment, trace the flow path from the faucet and identify all components from the P&ID. Draw the basic schematic and block diagram of the apparatus for PI control of the liquid level with the first and second control valve.

Now that you are familiar with the system you are ready for start-up. Hook up the ¼” air hose to the push-connect air line and the regulator on the cart. Check that all valves on the air-header are in the closed position. Turn the air on fully and adjust the regulator to supply 25 psi to the header. Next, screw the 1/2” OD polyvinyl hose to the water faucet and ensure a tight fit. Close the inlet ball valve before turning on the faucet. Turn the faucet handle for cold water fully open and check that no leaks are visible on the cart.

Open the LabVIEW 2013 program titled LevelControllerFinal.vi from the desktop. This is the graphical user interface and visual programming that performs the control operations. Plug all three USB cables into the computer and set the Output to COM11, Level transmitter to COM6 and Flow Transmitter to COM7. Click the Run button, the arrow in the top left of the toolbar to start the program. The default settings for both control valves is manual mode.

Open the green drain valve on the bottom of the cart and slowly open the inlet ball valve. Air that was trapped in the line will be forced out, so take care to open slowly until a smooth stream is seen entering the tank.

The first step is to become familiar with the process by manipulating the valve position manually and becoming familiar with the LabVIEW interface. The graphs display the controller output and sensor instantaneous sensor readings for the level transmitter and flow meter. To export this graph data to excel you simply right click on the graph, Export, export data to excel file. After each iteration of the experiment, or if other data is wished to be collected, the graphs must be cleared by right clicking-data operations-clear data, then right click again-data operations-make current value default. This will ensure the graph starts at time t=0 again.

Manually adjust the position of valve 1 and 2 until a steady state is seen on the level transmitter at 5-6”. This will be your Design Level of Operation (DLO) that all your control strategies are based on. Record the position of both valves and the inlet flow rate. Stop the program and clear the data, for you are about to perform the first “bump test”. Run the program for 2 minutes then manually set the first control valve to a position 1-3% away from the DLO. Use a stopwatch and visually record the level in the tank in addition to the automated data collection. Wait until the level in the tank reaches a new DLO. Export this data to Excel. Do this for both valves. Save the data from these experiments for next week

Day 2

This raw data will need to be filtered to make any useful calculations from it. Apply a filter method of your choice to acquire a reasonable looking step response. Mark when the valve position was changed and the time it took to first see a change in the level from its initial DLO. 1) This is the deadtime of the process, θd.

2) The change in the Process Variable (PV) over the change in the controller output (CO) is the process gain kp.

3) Next find where 63% of the change in process variable occurred and the time it took from the initial change in height to get there. This is the process time constant τp.

Alternatively you may perform a doublet test by manipulating the CO back to its original position after seeing a clear response in the PV.

You are now ready to write a First Order Plus Dead Time equation for the model from the form τp dPV(t)dt+PVt=kp*COt-θd

Using this FOPDT model you can create your PI controller equation

CO = controller output signal
CObias= controller bias or null value; set bybumpless transfer
e(t) = current controller error, defined as SP – PV
SP = set point
PV = measured process variable
Kc = controller gain, a tuning parameter
Ti = reset time, a tuning parameter

The tuning parameters and type of controller must then be chosen.

For this self-regulating process, PI control will be sufficient. The closed loop time constant τc can chosen between .1 and 10 for aggressive to conservative response.

You will notice that the same CO for one control valve has the reverse effect on the liquid level as the other. What is this called?

Input your tuning parameter for Kp and Ki on the front panel of the VI. Return the cart to steady state as in Day 1 and turn on Automated control. Implement disturbances by adding 1 liter of water to the vessel from a graduated cylinder and by closing the faucet 50%. Record the controller’s and process response and export to excel.

Day 3

Discuss the advantages and disadvantages of P-only and PI- control in respect to this process.

Compare the process gains you calculated from theoretical process models to the bump test results. Why are they different?

Investigate any other tuning methods learned in class and compare the results to Day 2.

Using the fundamentals you have learned in your Process Controls class, develop your own process model and calculate a theoretical gain. Is there a way to reject disturbance in the flow before the level in the tank changes? Of course there is, by cascading the dynamic relationship of the flow control valve to the draining of the tank. Develop a block diagram of this process using the flow control as a slave loop.