TEMPERATURE CONTROLLER SYSTEM

1. OBJECTIVE
To study the performance of various types of controllers used to control the temperature of an oven.

2. SYSTEM DESCRIPTION
Temperature control is one of the most common industrial control systems that are in operation. This
equipment is designed to expose the students to the intricacies of such a system in the 'friendly environment
of a laboratory, free from disturbances and uncertainties of plant prevalent in an actual process. The 'plant' to
be controlled is a specially designed oven having a short heating as well as cooling time. The temperature
time data may be obtained manually, thus avoiding expensive equipment like an X-Y recorder or a pen
recorder. A solid state temperature sensor converts the absolute temperature information to a proportional
electric signal. The reference and actual temperatures are indicated in degree Celsius on a switch selectable
digital display.
The controller unit compares the reference and the measured signals to generate the error. Controller options
available to the user consist of ON-OFF or relay with two hysteresis settings and combination of
proportional, derivative and integral blocks having independent coefficient settings. A block diagram of the
complete system is shown in Fig. 1.

3. BACKGROUND SUMMARY
The first step in the analysis of any control system is to derive its mathematical model. The various blocks
shown in Fig. 1 are now studied in detail and their mathematical descriptions are developed. This would
help in understanding the working of the complete system and also to implement control strategies.
3.1 The Plant (Oven)
Plant to be controlled is an electric oven, the temperature of which must adjust itself in accordance with the
reference or command. This is a thermal system which basically involves transfer of heat from one section
to another. In the present case we are interested in the transfer of heat from the heater coil to the oven and
the leakage of heat from the oven to the atmosphere. Such systems may be conveniently analysed in terms of
thermal resistance and capacitance as explained below. However, this analysis is not very accurate, since the
transfer of heat essentially takes place from every part of the oven thermal resistance and capacitance are
obviously distributed. The lumped parameter model described here is therefore only an approximation. For a
precise analysis, a distributed parameter model must be used. Another difficulty associated with temperature
control systems is that whereas the temperature rise is produced by energy input, which is controllable, the
temperature fall is due to heat loss which is uncontrollable and unpredictable. This implies that the oven will
have different time constants while heating and cooling. Again, these will depend on the ambient
temperature and the set point chosen. Such a system is therefore rather difficult to control. There are three
modes of heat transfer viz. conduction, convection and radiation. Heat transfer through radiation maybe
neglected in the present case since the temperatures involved are quite small. For conductive and
convective heat transfer
Θ=alpha*deltaT
where, rate of heat flow in Joule/sec.
delta T= temperature difference in °C
alpha =Constant
t= time in seconds
Under assumptions of linearity, the thermal resistance is defined as, R Temperature- difference/rate of heat
flow AT/1/a. This is analogous to electrical resistance defined by 1 V/R. In a similar manner thermal
capacitance of the mass is defined by
Θ = Cd(deltaT)/dt
which is analogous to the V-I relationship of a capacitor, namely IC dV/dt. In the case of heat,
C= Rate of heat flow/Rate of temperature change
The equation of an oven may now be written by combining the above two equations, implying that a part of
the heat input is used in increasing the temperature of the oven and the rest goes out as loss. Thus
Θ =CdT/dt+(1/R)XT,
with the initial condition T(t=0) Tamb. Now, taking Laplace transform with zero initial condition
T(s) /Θ(s) = R/ 1+SCR (1)
An analogous electrical network and block diagram may be drawn as in Fig.2 defined by the equation
I= CdV/dt + V/R
Eq.(1) is an extremely simplified representation of the thermal system under consideration and it gives rise
to a transfer function of the first order and type zero. Such a system should be easily controlled in the closed
loop. Difficulties are however faced in the system due to the following reasons:
(a) The temperature rise in response to the heat input is not instantaneous. A certain amount of time is
needed to transfer the heat by convection and conduction inside the oven. This requires a delay or
transportation lag term, exp(-sT₁), to be included in the transfer function, where T, is the time lag in seconds.
(b) Unlike the equivalent electrical circuit of Fig. 2, the heat input in the thermal system cannot have a
negative sign. This means that although the rate of temperature rise would depend on the heat input, the rate
of temperature fall would depend on thermal resistance R. The conventional analysis methods then become
inapplicable.
(c) Referring to the closed loop oven control system of Fig. 3, it may be seen that in the steady state the error
e, is given as

In this system, A cannot be increased excessively in an attempt to reduce error, since a large gain is likely to
lead to instability due to transportation lag. Also, every time (Tref-T) becomes negative, the heat input is cut
off and the oven must cool down slowly. The temperature T therefore oscillates around the nominal value.
3.2 Controller
Basic control actions commonly used in temperature control systems are

 ON-OFF or relay
 Proportional
 Proportional-Integral
• Proportional-Integral-Derivative
These are described below in some detail.
(a) ON-OFF or Relay type controllers, also referred to as two position controllers, consist of a simple and
inexpensive switch/relay and are, therefore, used very commonly in both industrial and domestic control
systems. Typical applications include air-conditioner and refrigerators, ovens, heaters with thermostat.
Solenoid operated two position valves are commonly used in hydraulic and pneumatic systems. The basic
input-output behaviour of this controller is shown in Fig. 4. The two positions of the controller are M1and
M2 and H is the hysteresis or differential gap.
The hysteresis is necessary, as it enables the controller output to remain at its present value till the input or
error has increased a little beyond zero. Hysteresis helps in avoiding too frequent switching of the controller,
although a large value results in greater errors. The response of a system with ON-OFF controller is shown
in Fig. 5. Describing function technique is a standard method for the analysis of non-linear systems, for
instance, one with an ON-OFF controller.
(b) Proportional controller is simply an amplifier of gain Kp which amplifies the error signal and passes it
to the actuator. The noise, drift and bias currents of this amplifier set the lower limit of the input signal
which may be handled reliably and therefore decide the minimum possible value of the error between the
input signal and output. Also the saturation characteristics of this amplifier sets the linear and non-linear
regions of its operation. A typical proportional controller may have an input-output characteristics as in Fig.
6. Such controller gives non-zero steady state error to step input for a type-0 system as indicated earlier. The
proportional (P) block in the system consists of a variable gain. amplifier having a maximum value, K pmax
of 20.
(c) Proportional-Integral (PI) controller: Mathematical equation of such a controller is given by
and a block diagram representation is shown in Fig. 7. It may be easily seen that this controller introduces a
pole at the origin, i.e. increases the system type number by unity. The steady state error of the system is
therefore reduced or eliminated. Qualitatively, any small error signal \mathfrak{c}(t) present in the system,
would get continuously integrated and generate actuator signal m(t) forcing the plant output to exactly
correspond to the reference input so that the error is zero. In practical systems, the error may not be exactly
zero due to imperfections in an electronic integrator caused by bias current needed, noise and drift present
and leakage of the integrator capacitor.
The integral (1) block in the present system is realised with a circuit shown in Fig.8 and has a transfer
function
The integral gain is therefore adjustable in the range 0 to 0.036 (approx.). Due to the tolerance of large
capacitance's, the value of K, is approximate.
(d) Proportional-Integral-Derivative (PID) controller: Mathematical equations governing the operation
of this controller is as

so that in the Laplace transform domain,

A simple analysis would show that the derivative block essentially increases the damping ratio of the system
and therefore improves the dynamic performance by reducing overshoot. The PID controller therefore helps
in reducing the steady state error with an improvement in the transient response.
The derivative (D) block in this system is realised with the circuit of Fig. 9. This has a transfer function
Gp(s) 19.97 s (approx.) (3)
The derivative gain is therefore adjustable in the range 0 to 20 approximately. Again, the approximation is
justified due to the higher tolerance in the values of large electrolytic capacitance's.
PID controller is one of the most widely used controller because of its simplicity. By adjusting its
coefficients Kp, Kp (or Tp) and K, (or Ti) the controller can be used with a variety of systems. The process
of setting the controller coefficients to suit a given plant is known as tuning. There are many methods of
'tuning' a PID controller. In the present experiment, the method of Ziegler-Nichol has been introduced which
is suitable for the oven control system, although better methods are available and may be attempted.
3.3 Temperature Measurement
The oven temperature can be sensed by a variety of transducers like thermistor, thermocouple, RTD and IC
temperature sensors. In the present setup, the maximum oven temperature is around 90°C which is well
within the operating range of IC temperature sensor like AD590. Further, these sensors are linear and have a
good sensitivity, viz. IµA/K.
Associated electronic circuits convert this output to 10mV/°C which may be easily measured by a DVM.
The time constant of the sensor has however been neglected in the analysis since it is insignificant compared
with the oven time constant.
4. EXPERIMENTAL WORK
A variety of experiments may be conducted with the help of this unit. The principal advantage of the unit is
that all power sources and metering are built-in and one needs only a watch to be able to note down the
temperature readings at precise time instants. After each run the oven has to be cooled to nearly the room
temperature, which may take about 15-20 minutes. This would limit the number of runs to about four in an
usual laboratory class. The experiments suggested could be completed in about 6-8 hours.

4.1 Identification of Oven Parameters

Plant identification is the first step before an attempt can be made to control it. In the present case, the oven
equations are obtained experimentally from its step response as outlined below. The procedure is as per
Ziegler-Nichol reaction curve method.
In the open-loop testing, the oven is driven through the P-amplifier set to a gain of 10. The input to this
amplifier is adjusted through reference potentiometer (the one next to switch S2). This input can be seen on
digital display, so that when you set 5.0°C, the input to Proportional amplifier is 50 mV (@ 10mV/°C) and
its output (which acts as input to driver circuit) is 0.5V (50mVx10).
 Keep switch S to 'WAIT, S₂ to 'SET' and open 'FEEDBACK' terminals. (refer panel drawing)
 Connect P output to the driver input and switch ON the unit.
 Set P potentiometer to 0.5 which gives K, 10. Adjust reference potentiometer to read 5.0 on the
DVM. This provides an input of 0.5 V to the driver.
 Put switch S₂ to the 'MEASURE' position and note down the room temperature.
 Put switch S₁ to the 'RUN' position and note temperature readings every 15 sec., till the temperature
becomes almost constant.
 Plot temperature-time curve on a graph paper. Referring to Fig. 10, calculate Ti and T₂ and hence
write the transfer function of the oven including its driver as G(s)=K *exp(-sT2)/(1+sT₁), with T in
°C.
4.2 ON-OFF Controller
 Keep switch St to 'WAIT position and allow the oven to cool to room temperature. Short
'FEEDBACK' terminals.
 Keep switch S₂ to the 'SET position and adjust reference potentiometer to the desired output
temperature, say 60.0°C, by seeing on the digital display.
 Connect R output to the driver input. Outputs of P, D and I must be disconnected from driver input.
Select 'HI' or 'LO' value of hysteresis. (First keep the hysteresis switch to 'LO').
  Switch S₂ to 'MEASURE' and S, to 'RUN' position. Read and record oven temperature every 15/30
sec., for about 20 minutes.
 Plot a graph between temperature and time and observe the oscillations (Fig. 15) in the steady state.
Note down the magnitude of oscillations.
 Repeat above steps with the 'HI' setting for hysteresis and observe the rise time, steady-state error
and percent overshoot.

4.3 Proportional Controller:

Ziegler and Nichols suggest the value of Ky for P-Controller as
Kp= (1/k)*(T1/T2)
 Starting with a cool oven, keep switch S₁ to 'WAIT position and connect P output to the driver input.
Keep R, D and I outputs disconnected. Short 'FEEDBACK' terminals.
 Set P potentiometer to the above calculated value of K P* keeping in mind that the maximum gain is
20. The measurement and interpretation of Kp and P-control potentiometer setting needs some
explanation here. The formula for Kp above is for an unity feedback system and has the dimension of
Volts/"C. In the present unit a temperature sensor having sensitivity of 10mV/°C (0.01V/°C) is used
between oven output and controller input. Thus, the Kp calculated above will need to be divided by
0.01 to obtain the P-control potentiometer setting. Kd and K_{1} have dimensions of sec. and sec¹
respectively hence do not require any further consideration. These values may be set directly on the
respective potentiometers.
 Select and set the desired temperature to say 60 deg * C
 Keep switch S_{1} to 'RUN' position and record temperature readings as before.
 Plot the observations on a linear graph paper and observe the rise time, steady- state error and
percent overshoot.

4.4 Proportional-Integral Controller

Ziegler and Nichols suggested the value of K_{P} and K_{1} for P-I controller as
Kp= (0.9/K)* T1/T2 1/K1=3.3*T2, giving K1 = 1/(3.3*T2)
 Starting with a cool oven, keep switch S_{1} to 'WAIT', connect P and I outputs to driver input and
disconnect R and D outputs. Short feedback terminals.
 Set P and I potentiometers to the above values of K_{p} and K_{I} respectively, keeping in mind
that the maximum value of Kp is 20 and that of K1 is 0.036.
 Select and set the desired temperature to say 60 deg * C
 Keep switch S_{1} to 'RUN' position and record temperature readings as before.
 Plot the response on a graph paper and observe the steady state error and percent overshoot.

4.5 Proportional-Integral-Derivative Controller

Ziegler and Nichols suggest the values of K_{p} K_{D} and K_{1} for this controller as
Kp = (1.2/K) * T1/T2 ; T1 = 1/K1 = 2T2 giving K1 = 1/(2T2)
Kd = Td = 0.5T2
 Starting with a cool oven, keep switch S_{1} to 'WAIT' position and connect P, D and I outputs to
driver input. Keep R output disconnected. Short feedback terminals.
  Set P, I and D potentiometers according to the above calculated values of K_{p} K_{1} and K_{D}
keeping in mind that the maximum values for these are 20, 0.036 and 23.5 respectively.
 Select and set the desired temperature, say 60 deg * C
 Switch S1 to 'RUN', and record temperature-time readings.
 Plot the response on a linear graph paper and observe the rise time, steady state error and percent
overshoot. (See Fig. 14)
 Compare the results of the various controller options.

4.6 Further Experimentation

The controller settings suggested by Ziegler and Nichols are not optimum. It is therefore possible to
experiment with other methods available in the literature or to attempt trial and error settings.
Students at the master's level may attempt to calculate theoretically the optimum values of Kp, Kd
and Ki based on some performance criterion and then verify them results on the setup. It may be
convenient to use a pen recorder or X-Y recorder for such experiments. A terminal has been
provided at the back of the unit for this purpose with a sensitivity of 10mV/°C.

Additional laboratory work may involve modification of the oven parameters and then repeating the
basic experiments. This may be done simply by putting thermal load into the oven, thus increasing its
thermal capacitance or by providing insulation to the oven thus increasing its thermal resistance.
 These may also act as disturbance inputs to the oven while it is operating under steady-state
conditions, and their effect may be studied.

5. TYPICAL RESULTS

(a) Open loop measurement: The constant K for oven plus driver controller is given by
K= (Final temperature Oven - Ambient temp)/ Input (Volts) = 87.2-34/0.5 = 53.2/0.5 = 106.4

From the graph between temperature and time Fig.11, the final oven temperature for a input of 0.5
volt is 87.2 deg * C. Hence, K = 53.2/0.5 = 106.4 deg * C / V .With reference to Fig.10, T1 and T2
as measured from the open-loop graph are: T1 = 312 sec : T2 =21sec. (Note that these values may
differ from unit to unit).

(b) Calculation for KP KI KD : The coefficient settings according to Ziegler and Nichols are
different for different types of control. The calculations for them are illustrated below.
(i) P Control: Kp=312 106.4 21 = 0.1396 V/°C With temperature-sensor sensitivity of 10mV /^ C
and maximum gain of P-amplifier as 20, actual Kp max. 0.1V/°C. Hence P-setting required for
proportional control is 70%. The Temperature vs Time plot is shown in Fig. 12.

(ii) P-I Control: Kp = (0.9/K) * T1/T2 = (0.9/106.4) * 312/21 = 0.1256

hence, P-setting required = 63%.
T1 = 3.3*t2 = 3.3 * 21 = 69.3 sec;
K1 = 1 / T1 = 1/69.3 =0.0144/sec K
K1-max . =1/41=0.0244, (see Eq. 2)
I-setting (0.0144/0.036)×100=0.40×100=40%
The Temperature vs Time plot is shown in Fig.13.
 (c) Results: Fig. 12-15 show the graph of temperature vs. time using P, PI, PID controller with above
coefficient settings, for a set temperature of 60 deg * C and also the realy control. A comparison of
the graphs with that obtained using control only should reveal the effectiveness of I and D controls in
reducing steady-state error and percentage overshoot.

6. LIMITATIONS OF THE SYSTEM

It must be appreciated that this is a purely experimental model designed for studying the
different control strategies available for controlling temperature. No effort has therefore been made
to optimise a particular method for the circuits involved, which would be possible and desirable in an
actual industrial controller. Nevertheless the salient features of the techniques have been brought out
clearly for an easy understanding.
The performance of the system is naturally limited by the imperfections of the components used.
These include the offset and drift in the operational amplifiers, leakage of the integrator and
differentiating capacitors, and temperature dependence of a number of components. As a
consequence, the steady-state error is not exactly zero inspite of Integral control in operation,
contrary to what would be expected from theoretical analysis. Also, the accuracy of the temperature
displayed on the panel may not be better than +5%.

Conclusion:-

In studying different controllers for oven temperature control, PID controllers stand out for their ability to
provide precise and stable regulation by combining proportional, integral, and derivative actions. While
proportional controllers are fast but leave steady-state errors, and PI controllers eliminate errors but may
respond slowly, PID controllers offer the best balance. On/off controllers, though simple, tend to cause
temperature oscillations, making them less suitable for precise control. Overall, PID controllers deliver the
most effective performance for maintaining consistent oven temperatures with minimal errors