International Journal of Latest Trends in Engineering and Technology

Vol.(9)Issue(2), pp.087-094
DOI: http://dx.doi.org/10.21172/1.92.15
e-ISSN:2278-621X

PIC 16F877A BASED IMPLEMENTATION OF FUZZY CONTROL

ALGORITHM FOR SPLIT AIR-CONDITIONING SYSTEM
C.B.Patil1, S. R. Potdar2 & R.R.Mudholkar3

Abstract- This paper reports the design and implementation of a microcontroller based fuzzy logic controller for Air
conditioning system. The proposed control strategy is planned to improve the performance of the Air conditioning system in
terms of energy consumption and comfort indoor temperature by Fuzzy Logic Control Algorithm (FLCA) that is implemented
in PIC 16F877A microcontroller through C-coding. The FLCA maintains the desired temperature value inside the prototype
room by varying the duty cycle of PWM signal generated on microcontroller pin through which the rotational speed of air
conditioning compressor is varied. In other words it controls the refrigeration cycle. The Fuzzy Logic Controller is compared
with ON/OFF controller and it is observed that there is significant improvement with respect to performance parameters.
Keywords – microcontroller, Air conditioning system, Fuzzy Logic Control Algorithm, PIC 16F877A, C-coding, PWM

1. INTRODUCTION
US Energy Information Administration (EIA) has reported that energy demand for space cooling is growing rapidly in India
and all around the world, because of rising prosperity of mankind. Concurrently due to global warming, cooling degree days
per year are increasing notably in the world, which further demands energy for space cooling. The current scenario reveals an
exponential increase in use of room air conditioning. Among the various air conditioning systems, split air-conditioning units
are generally used for small and medium scale offices and residential buildings.
Basically Air-Conditioning Systems are non-linear in which classical controllers such as ON/OFF and PID controllers are
used. ON/OFF controller has simple structure and low initial cost. However it is more expensive due to their low energy
efficiency. PID controller tends to be complex structure and initial cost is more. But, efficiency is more as compared to
ON/OFF Controller. However PID Controller is set to operate at designed thermal loads, while actual thermal loads are time
varying and depend on the environmental factors like outside weather conditions and number of occupants in the building.
One way to overwhelm this is to use of Fuzzy Logic Controller, because it has ability to tackle multi-parameters, besides non-
linear behavior of a system can be easily handled through fuzzy inference [1]. Many researchers are working on energy saving
by applying different strategies on different parts of air conditioning unit. According to A. Benamer and D. Clodic, up to 40%
energy saving is possible with variable speed compressor as compared to fixed speed compressor at lower heat loads [2]. Hu
and Huang presented water cooled air conditioner that uses cellulose pads as a filling material of cooling tower. They showed
that water cooled condenser results in decreasing the power consumption and improvement in COP significantly [3]. Amr O.
Elsayed investigated that a condenser air flow can reduce about 10% in compressor power consumption [4].

2. SYSTEM IMPLEMENTATION
A schematic diagram of experimental setup is as shown in figure 1 and schematic design of temperature controller in figure 2.

Figure 1. Schematic diagram of experimental setup

1
Department of Electronics, D. K. A. S. C. College, Ichalkaranji, Maharashtra, INDIA-416115, (Affiliated to Shivaji
University, Kolhapur, Maharashtra)
2
Department of Physics, Arts, Science and Commerce College, Rahata, Maharashtra, INDIA-423107
3
Department of Electronics, Shivaji University, Kolhapur, Maharashtra, INDIA-416004
 Pic 16f877a Based Implementation Of Fuzzy Control Algorithm For Split Air-Conditioning System 088

Figure 2. Schematic design of temperature controller

An isolated prototype room of dimension „2ft * 3ft * 4ft‟ made from thermocol with thickness 0.038m is considered inside
which an evaporator unit of split AC is mounted. Split AC has various components for vapor compression cycle: a compressor,
a condenser, an evaporator, metering device and fans. IC LM 35 sensor is used to sense the room temperature, which outputs an
analog voltage. The analog output of sensor is given to ADC of PIC micro-controller, where it is converted into digital signal
using scaling factor. The converted digital output is compared with set temperature value and accordingly an error signal is
generated. The error signal is given as input to Fuzzy Inference System (FIS) implemented in PIC 16F877A microcontroller
through C-code. FIS will then produce a crisp (count) value based on the control knowledge impersonated in the rule base.
Depending upon count value micro-controller generates a Pulse Width Modulated (PWM) signal. The PWM signal is converted
to proportional dc voltage in the range 0-10V with op-amp configured in non-inverting amplifier mode with gain of 2. The
output of op-amp amplifier is fed to VFD (Variable Frequency Drive) which drives the compressor accordingly. In turn the
evaporator temperature is changed and the system set in loop in maintaining the set temperature. The system simulink model in
which, the Fuzzy Logic Controller modulates the ambiance of Air conditioned prototype room is shown in figure 3[5].

Air-conditioner Prototype Room

Figure 3. MATLAB simulink model of plant having prototype room and air conditioner with fuzzy logic controller
 C.B.Patil, S. R. Potdar & R.R.Mudholkar 089

3. FUZZY LOGIC CONTROLLER

Basically two types of Fuzzy Inference Systems (FIS), „Mamdani‟ and „Takagi-Sugeno‟ are used in control applications. They
differ in compilation of output. Mamdani Fuzzy Controller has three stages of operations: fuzzification, fuzzy inference and
defuzzification. In present system, Mamdani Fuzzy Controller is being used for fuzzy inference.
The FLC employed has two input fuzzy variables: error (e) and rate of change of error (de), one output fuzzy variable: count
(u). The error is the difference between reference (set) value and current (measured) value of air temperature of room and rate of
change of error is the rate of change between current error and previous error in time interval of 5 seconds. The count means a
value in the range 0 to 255 that decides the duty cycle of PWM signal. The universe of discourse for (e) is -200C to +200C, for
(de) is -10C to +10C, for (u) is 0 to 255. The input variable error (e) is partitioned into three fuzzy sets viz. NEGATIVE (N),
ZERO (Z), POSITIVE (P), the input variable rate of change of error (de) is partitioned into three fuzzy sets viz. NEGATIVE
(N), ZERO (Z), POSITIVE (P) and output variable count (u) is partitioned into three fuzzy sets viz. LOW (L), MED (M), HIGH
(H) using triangular Membership functions. Fuzzy membership functions after tuning being applied are shown in figure 4, 5 and
6. The surface view of these variables is as shown in figure 7.

Figure 4. Membership functions for input variable „error‟

Figure 5. Membership functions for input variable „rate of change of error‟

Figure 6. Membership functions for output variable „count‟

 Pic 16f877a Based Implementation Of Fuzzy Control Algorithm For Split Air-Conditioning System 090

Figure 7. Surface view for input error (e) and rate of change of error (de) and output count(c)

3.1 Fuzzification-
In this process, the degree of membership of particular fuzzy set for the input variables is calculated using the following
formulae:
a) Triangular Membership Function-
This MF is defined as,

Alternative expression obtained by min and max operations is follows:

Triangle(x; a, b, c) = max(min((x-a)/(b-a), (c-x)/(c-b)),0)
Where the parameters „a‟, „b‟ and „c‟ are the x-coordinates of the three corners of the triangular MF.
b) Trapezoidal Membership Function -
This MF is defined as,

Alternative expression obtained by min and max operations is follows:

Trapezoid(x; a, b, c, d) = max (min ((x-a) / (b-a), 1, (d-x) / (d-c)), 0)
Where the parameters „a‟, „b‟, „c‟ and„d‟ are the x-coordinates of the four corners of the trapezoidal MF.

(a) (b)
Figure 8. Types of Membership functions (a) Triangular (b) Trapezoidal

The fuzzification process is illustrated as follows:

If error is -3 then the degree of memberships for input variable error (e) is calculated as shown in figure 9(a).

(a) (b)
Figure 9. Illustration of fuzzification process for (a) input error (-3) and (b) rate of change of error(-0.8)
 C.B.Patil, S. R. Potdar & R.R.Mudholkar 091
Degree of membership for NEGATIVE fuzzy set (d1) = μN(e) triangle = 0-(-3)/0-(-20) = 3/20 = 0.15
Degree of membership for ZERO fuzzy set (d2) = μZ(e)triangle = -3-(-5)/0-(-5) = 2/5 = 0.4

If rate of change of error is -0.8 then the degree of memberships for input variable rate of change of error (de) is calculated as
shown in figure 9(b).

Degree of membership for NEGATIVE fuzzy set (d3) = μN(de) triangle = 0-(-0.8)/0-(-1) = 0.8/1 = 0.8
Degree of membership for ZERO fuzzy set (d4) = μZ(de)triangle = -0.8-(-1)/0-(-1) = 0.2/1 = 0.2

3.2 Fuzzy inference-

In this process the degree of membership for each linguistic input term triggers the implication and derives a aggregated fuzzy
output set. This task can be accomplished by referring to the data base, knowledge base and reasoning process. Reasoning
process activated for priori formulated control laws relate the fuzzy input to fuzzy output from system behavioral knowledge as
follows –
Rule 1: IF (error is NEG) and (rate rate of change of error is NEG) THEN count is HIGH
Rule 2: IF (error is NEG) and (rate rate of change of error is ZERO) THEN count is HIGH
Rule 3: IF (error is NEG) and (rate rate of change of error is POS) THEN count is HIGH
Rule 4: IF (error is ZERO) and (rate rate of change of error is NEG) THEN count is MEDIUM
Rule 5: IF (error is ZERO) and (rate rate of change of error is ZERO) THEN count is MEDIUM
Rule 6: IF (error is ZERO) and (rate rate of change of error is POS) THEN count is LOW
Rule 7: IF (error is POS) and (rate rate of change of error is NEG) THEN count is LOW
Rule 8: IF (error is POS) and (rate rate of change of error is ZERO) THEN count is LOW
Rule 9: IF (error is POS) and (rate rate of change of error is POS) THEN count is LOW

Rule 1 implies that the frequency of PWM signal generated is high, which delivers high voltage to VFD that drives the
compressor with high speed of about 1150 rpm.
Rule 5 suggests that the frequency of PWM signal generated on PIC controller delivers a medium voltage, so that compressor
rotates with medium speed of about 575 rpm.
Rule 9 infers that the frequency of PWM signal generated is very small, which drives the compressor with speed of about 64
rpm.
rules are to be evaluated to compute the degree of satisfaction to MF (LOW, MEDIUM, HIGH) of the output (count) pertaining
to the degree of membership of the input variables. If error (e) = -3 and rate of change of error (de)= -0.8 only four rules get
fired. The fuzzy Inference process is illustrated as shown in figure 10.
Rule 1: IF (e is NEG) (degree of 0.15) and (de is NEG) (degree of 0.8) THEN count is HIGH (degree of 0.15) (AND)
Rule 2: IF (e is NEG) (degree of 0.15) and (de is ZERO) (degree of 0.2) THEN count is HIGH (degree of 0.15) (AND)
Rule 4: IF (e is ZERO) (degree of 0.4) and (de is NEG) (degree of 0.8) THEN count is MEDIUM (degree of 0.4) (AND)
Rule 5: IF (e is ZERO) (degree of 0.4) and (de is ZERO) (degree of 0.2) THEN count is MEDIUM (degree of 0.2)(AND)
Since it is two input system with AND operator, the MF of the output will have the degree which is less among the
memberships of input.

Figure 10. Illustration of fuzzy Inference for input error (-3) and rate of change of error (-0.8)
 Pic 16f877a Based Implementation Of Fuzzy Control Algorithm For Split Air-Conditioning System 092
3.3 Defuzzification:
In this process from the aggregated fuzzy output, a crisp or numeric output value is calculated. There are numerous variants of
the defuzzification procedure, such as Center of Gravity (COG), Centre of Sums (COS), Mean of maxima (MOM), First of
maxima, Last of maxima, Weighted Average Method (WA). In present study the Weighted Average Method is used which can
be represented mathematically as follows:

Where n = number of rules fired

= height of the rth clipped output fuzzy set.
= Peak values of the rth clipped output fuzzy set.
Referring to the figure 9 the resulting fuzzy output at the end of the inference is converted into crisp value as follows:
c  ( 0.15*2550.4*58) 111.72
0.150.4
The percentage PWM is calculated as follows:
%pwm ( cr*100)  ( 111.72*100) 43.81%
255 255

4. ‘C’ PROGRAM MODULE

// fuzzy implementation//
if((e<-20 ) && (de>=-1 && de<=0)) //CASE 1: e represents error and de represents error rate
{ cr=255;} // pwm with 100 % duty cycle
else if((e<-20 ) && (de>0 && de<=1)) //CASE 2
{ cr=255;} // pwm with 100 % duty cycle
else if((e>=-5 && e<=0) && (de>=-1 && de<=0)) //CASE 5
{d1=(0-e)/20; // μN(e)triangle
d2= (e+5)/5; // μZ(e)triangle
d3=(0-de)/1; // μN(de)triangle
d4=(de+1)/1; // μZ(de)triangle
if (d1<=d3)
{a1= d1;}
else
{a1= d3;}
if (d1<d4)
{ b1= d1;}
else
{b1= d4;}
if(a1>=b1)
{m3=a1;}
else
{m3=b1;}
if (d2<=d3)
{c1= d2;}
else
{c1= d3;}
if (d2<d4)
{e1= d2;}
else
{e1= d4;}
if(c1>=e1)
{m2=c1;}
else
{m2=e1;}
cr =((m2*58)+(m3*255))/(m2+m3);}//calculation of count (cr) by weighted average method of defuzzification
 C.B.Patil, S. R. Potdar & R.R.Mudholkar 093
5. RESULT AND DISCUSSION

Figure 11. simulated results for Room-temperature, DoM, Crisp value, PWM value on Real_PIC simulator

The C-code being executed in Real-PIC simulator for three different situations and the results are as shown in figure 11. It is
observed that when the room temperature is 30 0C which is above set point, the generated crisp value is 127.5, which produces
the PWM wave of 50% duty cycle. When the room temperature is 23 0C, the generated crisp value is 111.72, which produces
the PWM wave of 43.81% duty cycle. For room temperature of 16 0C which is below set point, the generated crisp value is
17.55, which produces the PWM wave of 6.8% duty cycle. This shows that duty cycle of PWM signals changes according to the
error and rate of change of error.
In order to observe the performance of Fuzzy Logic Controller embedded in PIC, a set temperature value of 20 0C was given
and the room temperature was logged. The data collected is plotted as shown in figure12. The MATLAB_Simulink model of
plant has been executed and the result plotted is shown in figure 12. The computer simulation and experimental results are
comparable which implicates the validity of the outcome. The little differences in parameter values are observed because of
thermal mass effect that causes time lag in experimental readings which is not considered in computer simulation.

Figure 12. Variation in Room _Temperature for different controllers

Also the ON-OFF controller embedded in PIC was executed for same set point value of 20 0C. The result plotted is shown in
figure 12. Because of continuous turn ON and OFF mechanism and overshooting the power consumption is increasing
significantly which further makes wear and tear of air conditioner parts noticeably. The comparison between Fuzzy controller
and ON_OFF controller results are tabulated in Table 1. It is observed that ON-OFF controller reaches the set value earlier than
Fuzzy controller but with considerable overshoots and undershoots.

Table -1 Performance Matrix for simulated and experimental results

Fuzzy controller(error and rate error) ON_OFF controller

Parameters experimental experimental
Simulation results
results results
Rise time 98 sec 150 sec 78 sec
0
Undershoot 0 C 0 0C 1.39 0C
Overshoot 1 0C 1.25 0C 2.3 0C
Settling time 480 sec 420 sec ----
Steady state error 0.02 0C 1.1 0C ----
 Pic 16f877a Based Implementation Of Fuzzy Control Algorithm For Split Air-Conditioning System 094
6. CONCLUSION
The fuzzy algorithm to control the air conditioner temperature has been successfully designed and implemented in 8-bit PIC
16F877A microcontroller. With few rules, a control is being accomplished in a stand-alone hardware. Hence the design
approach presented in this paper is cost effective. Further energy consumption is reduced greatly in fuzzy logic controller than
ON-OFF controller, addition to prolong the wear and tear of air conditioner. Hence there is significant improvement in the
performance of the Air conditioning system.

7. REFERENCES
[1] Rachana R. Mudholkar et al, “Genetic Tuning of Fuzzy Inference System for Furnace Temperature Controller”, (IJCSIT) International Journal of
Computer Science and Information Technologies, Vol. 6 (4) , 2015, 3496-3500.
[2] A.Benamer and D. Clodic, “Comparison of Energy Efficiency Between Variable and Fixed Speed Scroll Compressors in Refrigeration System”,
Proceedings of technological innovations in refrigeration in air conditioning and in the food industry into third millennium, June, 1999, pp.1-8.
[3] S.S. Hu and B.J. Huang, “Study of a High Efficiency Residential Split Water-Cooled Air Conditioner”, Applied Thermal Engineering, 25, 2005, pp.
1599–1613.
[4] Amr O. Elsayed1, Abdulrahman S. Hariri1 “Effect of Condenser Air Flow on the Performance of Split Air Conditioner”, World Renewable Energy
Congress 2011-Sweden, 8-13 May 2011.
[5] C.B.Patil et.al, “Thermal modeling of air-conditioned prototype room using Matlab- Simulink”, International Journal of Engineering Science and
Technology, Vol. 9(No.09), Sept. 2017, 901-907.