Abstract— The Calciner Unit plays an

important role in the modern cement industries as it is used for preheating the

raw materials like limestone which are fed into the kiln. The mathematical

model of the calciner unit is designed using System Identification technique for

the real time data obtained from the plant. A conventional PID controller has

been designed to control the temperature of the calciner unit. The parameter of

PID controller is tuned using Ziegler – Nichols tuning method. In order to

achieve optimum controller parameter a Self Tuning Fuzzy PID controller is developed. The

performance of the calciner unit has improved significantly compared to

conventional PID controller.

I. INTRODUCTION

Calciner temperature

control process is one of the most important processes in cement manufacturing.

It is used to maintain the raw mix texture, size of the mixture and perfect

blending of the raw material to produce more valuable clinker. Calciner unit is

used to preheat the raw mix sent into the kiln. The product obtained is

“clinker” (cement). Normal temperature of kiln is to be maintained at 800-960

°C and a normal coal feeding is 10-20 t/hr. There are four basic processes in

cement manufacturing. It starts with quarry where the raw material is extracted

and crushed. Then it will be sent to raw mill wherein the blending process

takes place (raw mix). The resultant from the above process was sent to the calciner

where the raw mix was preheated and fed into the kiln. The raw mix and fuel was

sent into the kiln. Clinker and exit gases come out. The clinker was sent to

finish mill, after which the size was reduced to obtain the final product

‘cement’. The basic schematic diagram of cement manufacturing plant is shown in

Fig.1.1.

Figure 1.1: Schematic diagram

of cement manufacturing plant

II. IDENTIFICATION OF SYSTEM

A. ANALYZING AND PROCESSING DATA

When preparing

data for identifying models, it was mandatory to specify information

such as input-output channel names, sampling time (10s). The toolbox helps to

attach this information to the data, which facilitates visualization of data,

domain conversion, and various preprocessing tasks. Measured data often has offsets, slow drifts,

outliers, missing values, and other anomalies. The toolbox removes such

anomalies by performing operations such as de-trending,

filtering, resampling, and reconstruction of missing data. The

toolbox can analyze the suitability of data for identification and provide

diagnostics on the persistence of excitation, existence of feedback loops, and

presence of nonlinearities. The toolbox estimates the impulse and

frequency

responses of the system directly from measured data. Using these

responses, system characteristics, such as dominant time constants, input

delays, and resonant frequencies can be analyzed. These characteristics can

also be used to configure the parametric models during estimation.

B. ESTIMATING MODEL

PARAMETERS

Parametric models, such as transfer functions or state-space models

use a small number of parameters to capture system dynamics. System

Identification Toolbox estimates model parameters and their uncertainties from

time-response and frequency-response data. These models can be analyzed using

time-response and frequency-response plots, such as step, impulse, bode plots,

and pole-zero maps.

C. VALIDATING RESULTS

System Identification

Toolbox helps to validate the accuracy of identified models using independent sets of measured data from

a real system. For a given set of input data, the toolbox computes the output

of the identified model and lets to compare that output with the measured

output from a real system. One can also view the prediction error and produce

time-response and frequency-response plots with confidence bounds to visualize

the effect of parameter uncertainties on model responses.

Figure

2.1: Shows the process of selecting the range for validation and estimation of

data.

D.

LINEAR MODEL IDENTIFICATION

System Identification

Toolbox lets to estimate multi-input, multi-output continuous or discrete-time transfer functions with a specified number of poles and zeros.

One can specify the transport delay or let the toolbox determine it

automatically. In this work, transfer function model was used for system

identification.

E. ESTIMATING TRANSFER

FUNCTION MODEL

Estimate continuous-time and discrete-time transfer functions and

low-order process models. Use the estimate models for analysis and control

design. Polynomial and

state-space

models can be identified using estimation routines provided in the

toolbox. These routines include autoregressive models (ARX, ARMAX), Box-Jenkins

models, Output-Error models, and state-space parameterizations. Estimation techniques

include maximum likelihood, prediction-error minimization schemes, and subspace

methods based on N4SID, CVA, and MOESP algorithms. A model of the noise affecting the observed

system can also be estimated. Figure 2.2 depicts the process of obtaining the

transfer function model.

Figure 2.2: Obtaining

transfer function model

F. ESTIMATING STATE-SPACE

MODEL

A state space

model is commonly used for representing a linear time invariant system. It describes

a system with a set of first order difference equation using inputs, outputs

and state variables. In the absence of the equation, a model of desired order

can be estimated for measured input, output data. The model was widely used in

modern control application for designing controllers and analyzing system

performance in the time domain and frequency domain. The models can be applied

to nonlinear system or system with a non-zero initial condition.

Figure

2.3: Obtaining state space model

III.

DESIGN OF PID CONTROLLER FOR CALCINER

A. PID CONTROLLER:

P-I-D controller has the optimum control dynamics including steady

state error, fast response, less oscillations and higher stability. The necessity

of using a derivative gain component in addition to the P-I-D controller is to

eliminate the overshoot and the oscillations occurring in the output response

of the system. One of the main advantages of the P-I-D controller was that it

can be used with higher order processes including more than single energy

storage.

From a mathematical viewpoint, the PID

control works to reduce the error e(t) to zero, where e(t) was the difference

between output response and the set point.

The

control response u(t) is given by:

u(t)=Kpe(t)+Ki?e(t)dt+Kd

de(t)/dt

where kp, ki, kd are scale factors for the

proportional, integral and differential terms respectively.

B. ZIEGLER – NICHOLS TUNING

METHOD:

The basic steps in Z-M method are

1. The value of Kd and Ki

were set to zero.

2. The value of Kp was slowly

increased such the sustained oscillation occurs (constant amplitude and

periodic).

3. The value of Kp at which

sustained oscillation occurs was ultimate gain Ku and the period of

oscillation was ultimate period Pu.

From the

calculated value of Ku and Pu, the parameters of PID

controller were calculated using the formula:

The table 3.1

shows the PID controller parameter tuned using Ziegler – Nichols method.

Table 3.1: PID controller tuning

parameters

Control type

Kp

Ki

Kd

PID

0.6*200=120

2/0.2=10

0.2/8=0.025

IV. DESIGN OF FUZZY CONTROLLER

Figure 4.1: General block diagram of fuzzy logic controller

A. FUZZY INFERENCE SYSTEM

A Fuzzy inference system (FIS)

was a system that uses fuzzy set theory to map inputs to outputs. There are two

types of FIS .They are mamdani and Takagi sugeno FIS. In this project there are

two inputs and three outputs. Therefore, mamdani type FIS was used in this

project.

i.

MAMDANI FIS

Mamdani FIS is

widely accepted since it can be applied for both MIMO, MISO systems whereas

sugeno can be implemented only for MISO systems. In mamdani, the membership

functions can be chosen even for outputs whereas it was not possible in sugeno

type. Hence mamdani FIS was used for our project.

ii.

DEVELOPMENT OF MAMDANI TYPE FIS

Calciner temperature in the cement

manufacturing process was developed using mamdani fuzzy model. It consists of

two inputs and three outputs. First input was error. Second input was rate of

change of error. The three outputs were Kp, Ki and Kd (i.e. controller gains).

Table 4.1:Rule table of fuzzy controller

B. MAMDANI FIS IMPLEMENTATION

FOR CALCINER TEMPERATURE CONTROL

Figure 4.2: Fuzzy logic toolbox

Figure 4.3: Membership function of inputs

Figure 4.4: Membership function of outputs

Figure 4.5: Rule viewer of mamdani FIS

Figure

4.6: Surface viewer of mamdani FIS

V. IMPLEMENTATION OF FUZZY

PID CONTROLLER

A. STRUCTURE OF FUZZY-PID

CONTROLLER

Self tuning fuzzy-PID controller means that

the three parameters Kp, Ki, and Kd of PID controller are tuned by using fuzzy

tuner. The coefficients of the conventional PID controller are not often

properly tuned for the non-linear plant with unpredictable parameter variations

.Hence, it was necessary to automatically tune the PID parameters.

Figure 5.1: Structure

of the self tuning fuzzy-PID controller

The error and the derivative of its error are sent to

the fuzzy controller. The PID parameter Kp, Ki and Kd is calculated according

to the rules in the fuzzy controller, at the same time, Kp was also refined by

P controller which was the immune PID controller, so the Kp, Ki and Kd can be

continuous updated according to error e(t) and its derivative de/dt.

VI. SIMULATION RESULTS AND

DISCUSSION

A. SERVO RESPONSE OF PID AND

FUZZY PID CONTROLLER

Simulation studies are carried out to

demonstrate the tracking capability of tuned PID controller and fuzzy PID

controller. The performance of process for tuned PID and fuzzy PID were shown

in figures 6.3 and 6.4 respectively. From the response, it was observed that

the calciner temperature follow the given set points and the servo response of

the PID and fuzzy PID were compared in the table 8.1.

Fig

6.1: Servo response of the PID controller

Fig

6.2: Servo response of the fuzzy PID controller

Table 6.1:

Comparison of performance indices of PID and FUZZY PID tuned controller for

servo response

CALCINER TEMPERATURE CONTROL USING

ISE

IAE

ITAE

PID CONTROLLER

1.559 e^(+05)

416.9

3975

FUZZY CONTROLLER

1.045 e^(+05)

279.3

2138

From the

responses, it was observed that the performance criterion such as ISE, IAE and

ITAE of Fuzzy PID controller was better compared to conventional PID

controller. It was also observed that fuzzy PID controller settles quickly than

PID controller response.

B.

SERVO WITH REGULATORY RESPONSE OF PID AND FUZZY PID CONTROLLER

Fig

6.5: Servo with regulatory response of the PID controller

Fig

6.6: Servo with regulatory response of the fuzzy PID controller

Table 6.2:

Comparison of performance indices of PID and FUZZY PID controller for servo

with regulatory response

CALCINER

TEMPERATURE CONTROL USING

ISE

IAE

ITAE

PID

CONTROLLER

1.605e^(+05)

622.8

9293

FUZZY

CONTROLLER

1.294

e^(+05)

410.9

4294

VII.

REAL TIME IMPLEMENTATION –CEMULATOR

Contrary to most cement process

simulators, ECS/CEMULATOR was developed on a full functional control systems

platform enabling the complete set of functions and features of a modern

control system environment for the users. Having a skilled team of operators

plays a crucial role in beneficial and safe operation of industrial plants.

Especially in the cement industry, with the significant high cost of

investment, practical knowledge and experience of plant operation have a direct

effect on production economy. Insufficient insight in process dynamics and

interactions, high stress factors in real time operation conditions, and lack

of adequate experience in utilizing the existing control system are typical

reasons for incorrect operator actions. The consequences of this may result in

low production quality, production interrupts, and equipment damage, in worst

case risk on human safety. The increasing demand on production sustainability

in the recent years has resulted in requirements of which the degree of

fulfillment is effected by the level of skills of plant operators and engineers.

A. REAL TIME RESPONSE OF THE PID CONTROLLER

Figure 7.1: Response of PID controller in real time

B. REAL TIME RESPONSE OF FUZZY PID CONTROLLER

Figure 7.2: Response of Fuzzy PID controller in

real time

Comparison of

performance indices of PID and FUZZY PID controller for the real time response

is shown in Fig. 7.1 and 7.2.

Table 7.1:

CALCINER

TEMPERATURE CONTROL USING

ISE

PID

CONTROLLER

18.4

FUZZY

CONTROLLER

16.4

From the table 7.1 it has been observed that Integral Square Error (ISE)

value of fuzzy PID controller is reduced as compared to PID controller.

VIII.

CONCLUSION

The main aim of the project was

to control the calciner temperature and to obtain a good quality clinker. The

transfer function model of calciner for the process has been derived using

system identification tool. The simulink model of calciner has been developed

in MATLAB using real time steady state values of Turkey power plant. The open

loop response of the process where observed and the interaction effect has been

studied. The parameters for PID were obtained using Ziegler – Nichols tuning. The

fuzzy rules were written using FAM table and the rules are inserted in the FIS

using mamdani method which is used to tune the PID. Thus Fuzzy PID controller

was implemented and then optimized values were obtained. It is observed that

the performance criteria namely the ISE, IAE, ITAE, and settling time in Fuzzy

PID controller is better than the PID controller. Also from the responses, it

has been observed that the proposed method has better tracking and faster

settling time.

IX. Appendix

DATA FROM REAL TIME CALCINER UNIT

S.NO

CALCINER TEMPERATURE

CALCINER COAL FEED

KILN TOTAL FEED

1

894.7916

9.6501

588.4775

2

894.7916

9.6401

589.4781

3

896.5278

9.6359

585.4742

4

898.9583

9.6276

588.4867

5

901.3889

9.6184

594.3333

6

904.1666

9.6096

590.6599

7

902.7778

9.6029

588.5881

8

900.6944

9.6033

590.9871

9

899.3055

9.6079

591.7212

10

901.3889

9.6074

589.3926

11

903.1249

9.6

585.8295

12

901.7361

9.5952

584.7019

13

900.6944

9.5972

586.1656

14

901.0416

9.5997

590.9084

15

903.1249

9.5979

590.3184

16

906.2499

9.5892

591.2415

17

904.8611

9.5817

590.2633

18

903.1249

9.5822

591.3748

19

902.7778

9.5847

591.8418

20

906.9444

9.5828

585.3685

References

1 Ashwani Kharola “A PID BASED ANFIS AND FUZZY

CONTROL OF INVERTED PENDULUM ON INCLINED PLANE”, International journal on smart

sensing and intelligent systems.

2 Ahmed M.Salch and Sohair F.Rezeta “FUZZY

LOGIC CONTROL OF AIR CONDITIONING SYSTEM IN RESIDENTIAL BUILDINGS”.

3 Elvira Lakovic and Damir Lotinac “AIRCRAFT

LANDING CONTROL USING FUZZY LOGIC AND NEURAL NETWORKS”

4 Gaddam Mallesham and Akula rajani “AUTOMATIC

TUNING OF PID CONTROLLER USING FUZZY LOGIC”, VIIIth international

conference on development and application system.

5 Jan Jantzen “TUNING OF FUZZY PID CONTROLLER”

,IEEE journal 2009

6 P.T.Rajeena Mol , B.Chinthamani, K.P.Kamini

and K.R.Sugashini “INTELLIGENT HOME HEATING SYSTEM” ,International Journal of

Research in Engineering and Technology

7 Salim Jyothi ohri “FUZZY BASED PID CONTROLLER

FOR SPEED CONTROL OF DC MOTOR” ,National institute of Technology.

8 Sankata B. Prusty, Umesh

C. Pati and Kamalakanta Mahapatra

“IMPLEMENTATION OF FUZZY-PID CONTROLLER TO LIQUID LEVEL SYSTEM USING LABVIEW”.

9 Somayeh Abdolzadeh and Seyed Mohammad Ali

Mohammadi “IMPLEMENTATION OF ADAPTIVE FUZZYCONTROLLER ON THE VARIABLE

SPEED WIND TURBINES IN COMPARISON WITH CONVENTIONAL METHODS”

10 S.Subramanian , N. Raghothaman and B.

Chellappa “FUZZY LOGIC CONTROL IN HIGH TEMPERATURE FURNACE” ,Recent Research in

Science and Technology ,2010.

11 Sudha Hatagar and S.V.Halse “THREE INPUT-ONE

OUTPUT FUZZY LOGIC CONTROL OF WASHING MACHINE”, International journal of

scientific research and Technology

12 Tushar Upalanchiwar and A.V.Sakhare “DESIGN

AND IMPLEMENTATION OF FUZZY-PID

CONTROLLER USING MATLAB”,International journal of research in computer and

communication technologies.

13 ZHANG Shengyi “STUDY OF FUZZY PID CONTROLLER

IN MATLAB FOR TWO PHASE HYBRID STEPPING”, IInd international conference on

system engineering and modeling.

14 Zulfatman and M. F. Rahmat “APPLICATION OF SELF-TUNING FUZZY PID

CONTROLLER ON INDUSTRIAL HYDRAULIC ACTUATOR USING SYSTEM IDENTIFICATION

APPROACH” ,International journal on smart sensing and intelligent systems, vol.

2, no. 2, june 2009.

15 Zuo Yusheng and Xu Qun “COMPARED WITH PID,

FUZZY AND FUZZY PID CONTROLLER”,IEEE journal.