Experiment No.

– 08 Date: 15 – 03 – 24
Conical Tank Level Process Trainer
AIM
To study the response of non-linear open loop control system in ON/OFF, P, PI, PD& PID controller.

THEORY
Conical Tank:
The level control of non-linear tanks (conical/spherical etc) is the immense challenge in process control and
it cannot be effectively controlled by means of conventional linear P+I+D controller. Conical tank is a highly
non-linear system. For complete drainage of fluids, a conical bottom cylindrical tank is used in some of the
process industries, where its nonlinearity might be at the bottom only. The drainage efficiency can be improved
further if the tank is fully conical. But continuous variation in the tank system makes it highly non-linear and
hence the liquid level control in such systems is difficult. A conical shaped tank system is mainly used in
Colloidal mills, Leaching extractions in pharmaceutical and chemical industries, food processing industries,
Petroleum industries, Molasses, Liquid feed and Liquid fertilizer storage, Chemical holding & mix tank,
Biodiesel processing and reactor tank. To avoid settlement and sludge in Storage and holding tanks, the conical
tanks are used.
Spherical tank:
The spherical tanks are predominantly used in chemical, petroleum refining, pulp and paper industries to store
and dispense the liquids with respect to quantity. The walls of the spherical tank act as a membrane to agitate
bending shear stresses acting on the tank. The spherical tank presents uncomplicated maintenance and efficient
as compared with that of conventional cylindrical tanks and cylindrical tank with conical bottom. In those
conventional tanks the wall stresses vary with direction to generate foam typically by agitation and hence
presents inefficient washing. The control of liquid level in spherical tank is difficult due to their nonlinear
dynamic behaviour and time domain specifications such as peak overshoot, settling time, peak time and rise
time. The controller calculates an error value as the difference between a desired set point and measured
process variable to optimize the error.

EXPERIMENTAL SETUP
The level process station was used to conduct the experiments and collect the data. The computer acts as a
controller. It consists of the software used to control the level process station. The setup consists of a process
tank, reservoir tank, control valve, I to P converter, level sensor and pneumatic signals from the compressor.

When the setup is Switched on, level sensor senses the actual level values initially, then signal is converted
to current signal in the range 4 to 20mA.This signal is then given computer through data acquisition cord.
Based on the values entered in the controller Settings and the set point the computer will take control action
the signal sent by the computer is taken to the station again through the cord. This signal is then converted to
pressure signal using I to P converter. Then the pressure signal acts on a control value which controls the
flow of water in to the tank there by controlling the level.
PROPORTIONAL CONTROL MODE
In this mode, over some ranges of errors about the setpoint, each value of error has a unique value of
controller output in one-to-one correspondence. The range of errors to cover the 0% to 100% controller
output is called proportional band, because the one-to-one correspondence exists only for errors in this
range. This mode can be expressed by

where, p0=controller output with no error (%)

INTEGRAL CONTROL MODE

where p(0) is the controller output when the integral action starts. The gain KI expresses how much
controller output in percent is needed for every percent-time accumulation of error.
It can be summarized as:
• If the error is zero, the output stays fixed at a value equal to what it was when the error went to zero.
• If the error is not zero, the output will begin to increase or decrease at a rate of KI percent/second for every
1% of error.
• The integral gain, KI, is often represented by the inverse, which is called the integral time, or the reset
action, TI =1/KI.
DERIVATIVE CONTROL MODE
Derivation controller action responds to the rate at which the error is changing— that is, the derivative of the
error. The equation for this mode is given by the expression

where the gain, KD, tells us by how much percent to change the controller output for every percent-per-
second rate of change of error. Derivative controller action is also called rate action and anticipatory control.

Procedure:
General Procedure
Hardware:
Note: Keep the appropriate positions of the three-way valves (V1, V2 and V3) to perform experiments on
either Conical Tank or on Spherical Tank.
1. Switch on electric supply. Switch on Mains.
2. Switch on Pump.
3. Adjust rotameter flow rate to 400 LPH.
4. Switch on compressor.

Software:
Note: The procedure for working with the software for both Conical and Spherical Tanks are same
and hence the explanations below are provided for Conical Tank Level Control.
1. Double click "ACE.exe" icon in desktop.
2. Select: Process Control Trainers ---> Non-Linear Process Control Trainers ---> Conical &
Spherical Tank Level Process Trainer (ACE 304) ---> Conical Tank Level Control ---> Open
Loop, On/Off, PID Control & Tuning / Gain Scheduling Control / Fuzzy Logic Control

On/Off Control
1. Select COM Port and Select On / Off Control
2. Change Hysteresis value to 5%. (Range 1-10%)
3. Change the values of the set point and observe the On-Off control operation.
Observations:
• Observe that if process value exceeds the set point and increases above the value of (0.5x Hysteresis),
control valve is fully closed and if process value decreases below the set point by (0.5 x Hysteresis), the
control valve opens fully i.e. controller operates like on/off switch.
P Mode
1. Select COM Port and Select P Mode
2. Keep the set point to 50%. Allow the process to stabilize.
3. Change output mode to Manual. Adjust output value so as to match the process value with set point and
apply this output value as bias value to the controller.
4. Switch the controller to Auto mode. Adjust the proportional band to 100%.
5. Apply step change to set point (say 50% to 60%) and observe the response.
7. Decrease proportional band to half of the previous value. With each decrease, obtain a new response of
the step change. Ensure that the set point changes are around the same operating point.
8. Using trial and error approach, find a value of proportional band so that the response to a step change has
at most one overshoot and one undershoot.
Observations:
• Observe steady state error decreases as proportional band decreases.
• Observe the effect of very low proportional band values (system works in oscillatory mode).
• Observe the response of the system to load change. Load change can be given by slightly opening the drain
valve for a shorter duration and retain the same position.
• Apply a setpoint change (say 50% to 60%) with the proportional band obtained using trial and error
approach. Plot the responses.
PI Mode
1. Select COM Port and Select PI Mode
2. Keep the set point to 50%. Allow the process to stabilize.
3. Set the proportional band estimated from Proportional control (P only) experiment and with integral time
= 50.
4. Decrease the integral time by half the previous value.
5. Apply step change to set point (say 50% to 70%) and observe the response.
6. Keep the set point to 50%. Allow the process to stabilize.
7. Decrease the integral time again half of the previous value and observe the process response for step
change.
8. Repeat the procedure until a satisfactory control is obtained with one overshoot and one undershoot and
with very less offset.
Observations:
• Observe the effect of reducing integral time on the response of the process. Also note that the process may
show very less or null offset as effect of integral action but with a sluggish response.
• Apply a setpoint change (say 50% to 70%) with the integral value obtained using trial and error approach.
Observe the responses and the time taken to reach steady state.

PD Mode
1. Select COM Port and Select PD Mode
2. Keep the set point to 50%. Allow the process to stabilize.
3. Set the proportional band estimated from Proportional control (P only) experiment and with derivative
time=1.
4. Allow the process to reach at steady state.
5. Increase the derivative time by 1 sec.
6. Apply step change to set point (say 50% to 60%) and observe the response.
7. Increase the derivative time gradually (1 to 2 and 2 to 3) and observe the process response for step
change.
Observations:
• Observe the effect of increasing derivative time. Also note that the process may show offset
as effect of integral action is cut off.
PID Mode
1. Select COM Port and Select PID Mode
2. Keep the set point to 50%.
3. Change the proportional band to the value that estimated in proportional controller.
4. Set integral time and derivative time based on the responses in previous experiments.
5. Allow the process to stabilize.
6. Apply step change to set point (say 50% to 60%) and observe the response.
7. Change the proportional band, integral time, derivative time as per the requirement and observe the
response of the process for step change for each change in setting.
Observations:
• Apply a setpoint change (say 50% to 60%).
• Adjust the P, I and D values using trial and error approach to obtain a better response.
• Achieve a step response with at least one overshoot and one undershoot.
• Plot the response.
• Compare the response with previous experiment responses.
• Vary the setpoint to any value in the operating range (Say 20% to 80% or 10% to 90%) and observe the
response.
• Compare the steady state response of the PID controller with P, PI and PD controller obtained in the
previous experiments.
Graphs
RESULT
The response of all the controllers on controlling the level in the conical tank is analysedand the graphs of the
set point v/s time and tank level v/s time is plotted and analysed. It isobserved that PID controller gives the
fastest response and is desired the most to use for levelcontrol. On-off control gave an oscillating response
and P controller gave the slowest response. PDcontroller gave fast response but had a significant offset. PI
controller gave a fairly faster response with no offset.

INFERENCE
All the controllers showed the predicted response. While PD controller showed offset thePI, P and PID
controller eliminated offset and made the response accurate. All the controllers servedifferent purpose in the
industry and each of them has their own advantages and disadvantages.
P Controller is very easy to implement as they rely on the current error between the desiredsetpoint and the
measured value. But they have a constant difference between the setpoint and theactual value. They are used
in liquid level control in tanks and temperature control in ovens where small offset is not tolerable.
PD Controller is very helpful in overshoot reduction and contribute to system stability. But it’sincreased
offset and low robustness stand for its main disadvantage. It is used in motor control andautomated systems.
PI controller very fast responding and it also helps in eliminating offset. But its increased complexity and
low robustness while handling nonlinear systems or rapidly changing conditions stand for its main
disadvantage.
PID Controller has better control performance and faster response. In the equation the integral term
eliminates offset and the differential term helps to dampen the system oscillations. And it is aversatile
system and can be used in various fields. But it is very complex to tune and theoveraggressive differential
action may lead to sluggish response. And it assumes a linear modelwhich is not true for all systems.