OBJECTIVES

1. To identify the important components of the air temperature control system.

INTRODUCTION & THEORY

Temperature is one of the process variables that is important as it provides the critical
condition for many industrial processes such as combustion, fermentation and crystallization.
Poor temperature control can cause major problems on safety, quality and productivity. In this
experiment, the equipment used to control air temperature is model AT922. This model uses
air to simulate a gas or a vapor temperature process which has the characteristic of multi
capacity lag or also known as dead time process. This model consists of a forced convective
annulus electric air heater to heat the incoming process air from an external compressor
supply system, a pressure regulator, a resistance temperature controller (RTD), a type K
thermocouple, an ON/OFF temperature controller and a PID temperature controller that is
used to maintain the air temperature at the heater exit.

PID controller consists of three actions which are proportional, integral and derivatives.
This controller is a control loop feedback mechanism that is widely used in industrial control
systems. The proportional action produces an output value that is proportional to the current
error value while the integral actions accelerates the movement of the process towards set-
point and eliminates the residual steady state error. Derivative action takes place by predicting
the system behaviour to improve settling time and stability of the system (Astrom, 2002). Other
than that, an ON-OFF controller is also used to control air heater process. The ON-OFF control
mode is the simplest control scheme as it only involves either a minimum, 0% or a maximum,
100% output signal from the controller based on the sign of the measured error. The controller
will switch its output only when the temperature crosses the setpoint. Dead band is usually
available with most controllers to reduce the frequency of oscillation (Process Control -
Temperature, n.d.).

Apart from that, PID tuning involves in selecting the process specific values for a range
of interacting parameters to achieve optimal performance. According to Astrom (2002), tuning
a PID controller careful adjustment of proportional gain, integral time and derivative time is
required to properly address the control requirements of the process. Some of the tuning
methods includes manual tuning, Ziegler-Nichols and Cohen-Coon. Ziegler-Nichols method is
a heuristic method and consists of two types: open loop and closed loop tuning method.

Open loop tuning method was developed by John G. Ziegler and Nathaniel B. Nichols
in 1942. P, I and D tuning constant is required for the control of the process based on the
open-loop response of the process to a step change in the set point. Based on the response
curve of the controlled variable, the process gain Kp, dead time td, and time constant 𝜏 must
be determined to calculate the value of parameter K. For non-interacting process, K can be
calculated by using equation as shown below:

𝜏
𝜅=| |
𝑡𝑑 𝐾𝑝

Figure 1 below shows control parameters for the open-loop Ziegler-Nichols tuning method.
This formula is used to calculate the tuning constant of the controller based the value of
parameter K (Process Control - Temperature, n.d.).

Figure 1: Control parameters for the open-loop Ziegler-Nichols tuning method

The objectives of this experiment were to control air heater process using PID
controller, to tune the PID controller by Ziegler Nichols open loop method and to control heater
process using the ON-OFF controller. This experiment consists of three part which was first
to determine the parameters of the transfer function G1 relate with the air temperature to the
electrical power input. Second was to determine the parameters of the transfer function G2
relating the air temperature to the air flow rate (disturbance or load variable). Lastly is control
of air heater process using PID Controller, there are two trials value for this part and the load
of disturbance is applied and have set point change.
 The experiment started by switch on the selector to panel SCADA position and make
sure the heater was switched off. Manual valve MV90A and heater by pass valve was fully
shut. Then MV90A inlet valve was fully opened so that the air flow rate FI90 reads at 45 nm3/hr
the RCD button is then switch on to record . The first part is started which to determine the
parameters of the transfer function G1 relating the air temperature to the electrical power input.
We then observed the response of RED PEN (TE91 /TIT91) and GREEN pen (TE92/TIT92).
The Equipment was let to have a stable reading of both green and red pen. Average flow rate
recorded is 44 nm3 /hr after that with TIC91 in Manual mode set MV to 30% the heater is then
switched on and the recorded chart is marked.

The next part of the experiment was about to determine the parameters of the transfer
function G2 relating the air temperature to the air flow rate (disturbance or load variable).The
system is ensure to be in a steady state condition. The vent VT is closed partially by about
two turns to disturb the air flow. The percentage of change in the air flow rate was recorded to
be decrease by around 10%. The heater is then switch off and recorder is turn off. By the
response on the chart paper the dead time can be calculated by drawing the steepest tangent.
The dead time in time units is given by:

(𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑖𝑛 𝑚𝑚 𝑥 3600𝑠𝑒𝑐)
Dead time = 𝑅𝑒𝑐𝑜𝑟𝑑 𝑐ℎ𝑎𝑟𝑡 𝑠𝑝𝑒𝑒𝑑,500 𝑚𝑚/𝐻𝑟

(5𝑚𝑚 𝑥 3600𝑠𝑒𝑐)
= = 36 secs
𝑅𝑒𝑐𝑜𝑟𝑑 𝑐ℎ𝑎𝑟𝑡 𝑠𝑝𝑒𝑒𝑑,500 𝑚𝑚/𝐻𝑟

While the process gain was as follows;

max 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑎𝑖𝑟 𝑡𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒

𝐾 = 1.64
The constant time,Ʈ can be calculated as follow:

Maximum change in the air temperature at the exit = 1980C – 1560C = 420C

𝑅𝑅
Ʈ=
max 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑎𝑖𝑟 𝑡𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒

102
Ʈ=
42

Ʈ = 2.429

The last part of this experiment Control of Air Heater Process Using PID Controller, for
this part the process supplied, and flow rate follow as set in section 3.3. The SSW was set to
1 so that TIC91 is used. This step makes the PID 2, Loop 2 disabled. The PID values in
controller TIC91/TIC92 is accessed and the following trial values as follows;

Trial 1 Trial 2
PB 10% 10%
TI 100s 22s
TD 25s 5s
SV 40C 80C
Figure 2: Trial 1&Trial 2

TIC91 was set to auto mode (A). Heater was switched on. The control response of the
heater surface temperature and heated air temperature are observed, the set point obtained
is 40C. As we can see the temperature drop of 420C. It is proven that the heated air
temperature overshoots the set point even when the heater surface temperature started
decreasing.

The last step was to apply two set of disturbance. For load disturbance the controller
is switched to manual mode and the MV was increase by 20%, then quickly switched back to
auto mode (A). We could observe a slight decrease in temperature by about 10C. The next set
of disturbance is about set point change. The air temperature is step increase by 100C to new
set point of 73.30C. The heater is then switch off and air flows is let to cool down the heater to
room temperature as a shutdown procedure.
CONCLUSIONS

To conclude, as the disturbance was increased, the temperature will increases, where
the manipulated are also increased from 30% to 45%. A tangent line is drawn from the cross
point of tangent until the baseline of set point. Based on the tangent being drawn, value of
time and rate of responses value are determined. By using the formula, value of dead time
was obtained which is 36 seconds, and the responses rate was 5 mm. From that, value of
process gain (K) and time constant (T) is obtained which K is 1.64 and Ʈ was 2.429, thus the
1.64
Parameter for First Order with Time Delay (FOTD) was G(s) = e−28.8 s. Next, second
2.429𝑠+1

experiment is conducted for two trials to set two differences Set-point Value (SV). For first trial,
the SV was set up to 40⁰C and it increased the MV value from 17.7% to 37.7% as the load
disturbance are set to 20% manually. Meanwhile, the temperature increase form 42⁰C to 40⁰C
before return to steady state. As conclusion, the experiment was said to be successful as the
objectives were achieved.

When conducting this experiment, there were few errors might happen leading to
inaccurate result obtained throughout the end of this experiment. The error that have been
detected such as the control valve that were supposedly closed or opened were not tight
closed or opened. This error contributed slightly in inaccurate graph obtained. In order to
prevent this error from occurring in the future, all the control valve must be checked at least
twice before stating the experiment. Besides, some of the graph did not show stable result.
This might due to the less time given for the system to stable. In order to minimise this error,
the systems need to leave for a while after introducing any disturbances or implying any
changes to the systems before taking any future action.

REFERENCES

Astrom, K. J. (2002). PID Control. https://doi.org/10.5739/isfp.2002.143

Process Control - Temperature. (n.d.).

Kharagpur. (n.d.). Module 6 Actuators. Version 2 EE IIT, 1–11.

Practical Control Valve Sizing , Selection and Maintenance (n.d.).