NeuralNetwork

CONTENTS

Ø Introduction to project.

v What is project?

v Our project neural network.

v Evolution.

v Why we use neural network.

v Controllable and uncontrollable parameters

Ø Selected manufacturing process.

v What is cement?

v Cement manufacturing process

Ø Quality characteristics.

v Specification.

v Defective.

Ø Artificial neural network.

v Introduction to ANN.

v Background.

v Models.

v Network in ANN.

v Application.

v Types of NN.

v Basic concept of neurons working.

v NN construction.

v Limitations of ANN.

Ø Cement manufacturing process in ACC.

Ø Introduction to matlab

v About matlab.

v Basics of matlab.

v Neural network toolbox in matlab.

v Structure of NN tool box.

Ø Blast furnace

v Iron making in blast furnace

v Raw materials section

§ Ore yard

§ Stock house

§ Coke charging system

§ Hoist house

v Furnace section

v Auxiliary section

v Controllable parameters …….

v Uncontrollable parameters …….

Ø INTRODUCTION TO PROJECT

v WHAT IS PROJECT: -

A technical student is expected to do some experimentation and research work on the subject, which he had learnt, at the classes during the course of his studies . Such an effort when well organized with a definite aim or purpose is called a PROJECT.

The object of a project is to evolve technical thinking ,analyse the problem ,

present the findings to the Professors and above all make a logical Engineer .

Through a project student displays his spirit of inquisitiveness , creativity , ability to understand a problem and analytical ways of solving the problem .

v OUR PROJECT :- PROCESS CONTROL IN INDUSTRIES USING NEURAL NETWORK

The term neural network is made by combining the words neural plus network. The word neural is related to brain and network implies some form of interconnection . So neural network implies interconnection between brain element .

v EVOLUTION

The term neural network was marked in year 1985 .

The reason for evolution of neural network is based on what we do (i.e..writing something , performing some gesture etc ) is due to the brain effect which is sensitive to information generated to mind . Brain contains many neurons which are interconnected forms the network .This principle is been given to us by god .

In neural network we are trying to replicate the priciples of brain to the maximum extent possible .

v WHY WE USE NEURAL NETWORK

Applying neural network in industrial process,implies,removing the complexities

In the process . Complexities can be explained by simple example as:-

General equation for line can be writtern as:

Y=mx+c

In complex form :

Y=m1x1+m2x2+c

Even in more complex form:

Y=m1x1+m2x2+m3x3+c

And so on.

We would apply neural network in the form of SIX-SIGMA in industrial process . In simple words , SIX-SIGMA can be explained as minimum rejections (say two in ten lakhs products made) .

v CONTROLLABLE AND UNCONTROLLABLE PARAMETERS

Controllable parameters are the parameters which are within our control.

For example : Temperature, pressure, flowrate, etc

Are the controllable parameters .

Uncontrollable

Parameters are the parameters in which we have no control . For example : Limestone contents (say SiO2,MgO,etc), which we receive from the minies ,

we have to use as such .

Ø SELECTED MANUFACTURING PROCESS :- CEMENT

v What is cement?

Cement is a binder, a substance which sets and hardens independently, and can bind other materials together. Most important cements are hydraulic cements, materials which set and harden after combining with water, as a result of chemical reactions with the mixing water and, after hardening, retain strength and stability even under water.

The most important use of cement is the production of mortar and concrete - the bonding of natural or artificial aggregates to form a strong building material which is durable in the face of normal environmental effects. By far the most common and most important hydraulic cement in modern construction is Portland cement, which is made primarily from limestone, certain clay minerals, and gypsum, in a high temperature process that drives off carbon dioxide and chemically combines the primary ingredients into new compounds. The name "cement" goes back to the Romans who used the term "opus caementitium" to describe masonry which resembled concrete and was made from crushed rock with burnt lime as binder. The volcanic ash and pulverized brick additives which where added to the burnt lime to obtain a hydraulic binder were later referred to as cementum, cimentum, cäment and cement.

v CEMENT MANUFACTURING PROCESS :-

***FLOW DIAGRAM OF CEMENT MANUFACTURING PROCESS AT
ACC JAMUL. ***

Raw coal

Clinker Storage Gantry

KILN

Primary and secondary crusher

Raw Mill

Lime stone Gantry

Cement Silos

PACKING HOUSE

CDS / Market

The cement manufacturing process is basically divided into five department .

v Crusher department.

v Raw mill.

v Kiln department.

v Cement mill.

v Packing house.

Raw material used in cement manufacturing are:-

v Lime stone.

v Sand.

v Clay.

v Shale.

v Iron ore.

v By products.

In this lime stone is the major product.

By products are the raw material which replaces natural raw materials in achieving sustainable development.

Lime stone contents and its composition:-

Lime (CaO):-(50-60)%

Silica (SiO2):-(20-25)%

Alumina (Al2O3) :- (5-10)%

Ferric oxide(Fe2O3) :- (1-2)%

Magnesium oxide (MgO):- (2-3)%

v CRUSHER DEPARTMENT :-

Limestone in the mines are drilled and blasted to a size of about 1.5 mts in diameter . This lime stone is brought to the crusher department where it is crushed to a size of 100 mm .

v RAW MILL AND RAW GRINDING:-

Raw grinding in raw mill is done by two processes.

· Wet process.

· Dry process.

In wet process, each raw material is proportion to meet desired chemical composition and fed to a rotating ball mill with water .Raw material are ground to a size where majority of materials are less than 75 microns.

Materials exiting the mill are called slurry and have flow ability characteristics. This slurry is pumped to blending tanks, and homogenized to insure the chemical composition of slurry is correct. Following the homogenization process the slurry is stored in tanks until required.

In dry process, each raw material is proportioned to meet the desired chemical composition and fed to either rotating ball mill or vertical roller mill .The raw material are dried with waste process gases and ground to a size where the majority of materials are less than 75 microns. The dry materials exiting either type of mill are called “Kiln feed”. The kiln feed is pneumatically blended to insure the chemical composition of the kiln feed is well homogenized and then stored in silos until required.

v KILN DEPARTMENT (PYRO-PROCESSING )

Whether the process is wet or dry, the same chemical reaction takes place. Basic chemical reaction are evaporating all moisture, calcining the limestone to produce free calcium oxide, and reacting calcium oxide with minor materials. This results in a final black nodular product known as “CLINKER” which has the desired hydraulic properties.

In wet process, the slurry is fed to a rotary kiln, which can be from 3.0m to 5.0m in diameter and from 120.0m to 165.0m in length. The rotating kiln is made of steel and lined with special refractory materials to protect it from high process temperature. Process temperature can reach as high as 1450 degree centigrade during the clinker making process..

In dry process, kiln fed is fed to a pre-heater tower, is discharged to a rotary kiln which can have the same diameter as a wet process kiln but the length is much shorter at approximate 450mts .The pre-heater tower and rotary kiln are made of steel and lined with special refractory materials to protect it from high process temperature.

Regardless to the process, the rotary kiln is fixed with an intense flame, produced by burning coal , coke ,oil, gas or waste fuel . Pre-heater tower can be equipped with firing as well.

The rotary kiln discharges the red hot clinker under the intense flame into a clinker cooler .The clinked cooler recovers heat from the clinker and returns the heat to the pyro-processing system this reduces fuel consumption and improving energy efficiency. Clinker leaving the clinker cooler is at a temperature conductive to being handled on standard conveying equipment.

v TEMPERATURE RANGE AT VARIOUS ZONES IN ROTARY KILN

§ Zone 1:-upper part of kiln .

Temp range :- (1000-1100 )Kelvin.

Charge losses all its water due to evaporation by hot gas.

§ Zone 2:-Middle part of kiln .

Temp range :- (1100-1200)Kelvin.

Lime stone decomposes to form CaO and Co2.

§ Zone 3:-Lower part of kiln .

Temp range :- (1770-1870)Kelvin .

As charges reaches here , chemical composition takes place between lime , alumina and silica to form calcium silicate [2CaO.SiO2 ; 3CaO.SiO2] and calcium aluminate [2CaO.Al2O3 ; 3CaO.Al2O3] .

v CEMENT MILL – (FINISH GRINDING &DISTRIBUTION) :-

The black nodular clinker is stored on site in silos or clinker domes until needed for cement production. Clinker, gypsum [CaSO4. 2H2O] , and other process additives are ground together in ball mills to form the final cement product. Fineness of final product, amount of gypsum added and the amount of process addition added are all varied to develop a desired quality of cement.

Clinker, along with additives is ground in a cement mill. The output of a cement mill is the final product (i.e.. CEMENT) .In all there is a

lying horizontal which contains metallic balls and as it rotates, the crushing action of the balls help in grinding the clinker to fine powder. The term bag house is applied to large filters containing a number of tubular bags mounted in a usually rectangular casing .The dust laden air is drawn through them by suction .The bag house is bused to remove dusty particle from discharge of different equipment such as cement mill, coal mill and kiln. In a bag house system discharge gas containing dusty particle is passed through a series of bags made of strong fabrics .

Important compounds present in cement are dicalcium silicate (Ca2SiO4 – 26%), tricalcium silicate (Ca3SiO3 – 51%) and tricalcium aluminate

(Ca3Al2O6 – 11%).

Strength of cement is due to the formation of - Si - O - Si – O - bond .

Ø QUALITY CHARACTERISTICS :-

Quality characteristics are the characteristics which are required to maintain the quality of the product for the end consumers.

Some of the characteristics are :

· Specification.

· Defectiveness.

v SPECIFICATION

It contains the list of essential characteristics and their tolerances .

§ Standard specification or normal specification.

§ Customer specification.

§ Company specification.

§ STANDARD SPECIFICATION

Simplified by the growth of standard for materials, components, processes, test, products etc.

BIS – Technical committee working within the framework of BIS for standardizing the specification of consumers as well as producer’s product.

§ CONSUMER SPECIFICATION

When BIS specification are not available or not suitable for a particular customer needs, the customer provides the specification to suit his particular needs and the manufacturer may agree and produce the product as per the specifications given by customer.

§ COMPANY SPECIFICATION

Where a company manufactures products to its own specification due to varied constraints and the customer accepts them, the specification may be called company specification.

v PRODUCT AND MATERIAL SPECIFICATION

§ SPECIFICATION IDENTIFICATION

This includes, the name of the product, a number which serves as a shorthand description, date, and issue number etc.

§ PRODUCT DESCRIPTION

This describes the product completely , utilizing shorthand designations.

Example :- make, size, grade, number of components and their description.

v DEFECTIVE :-

An item is said to be defective if it fails to confirm to the specification in any of characteristics . Each characteristic that does not meet the specification is a defect . An item is defective if it contains at least one defect .

According to their seriousness the defects can be classified as :-

1. Class A defect – very serious (critical):-

a. These defects will render product totally unfit for service , makes the products useless , unsalable .

b. Will cause-operating failure of the product in series, which can’t be readily, corrected example: open induction coil , transmitter without carbon , etc.

c. Liable to cause personnel injury or damage to the property .

2. Class B defects –Serious (major):-

a. These defects will probably , but not surely cause operating failure of the unit in service .

b. Will surely cause adjustment failure operation below standard etc .

c. Will surely cause increased maintenance

or decreased life.

3.Class C defects-Moderating serious :-

a.These defects will probably cause operating

failure of the unit in serious .

b. Likely to cause trouble of nature less

serious than operating failure .

c.Likely to cause increased maintenance or

decreased life .

d.Major defects of appearance , finish or workmanship .

4.Class D defects – not serious (minor) :-

a.These will not cause operating failure of the

unit in service.

b.It includes minor defects of appearance ,

finish or work man ship .

v PURPOSE OF SPECIFICATION :-

A specification is a defination of a design . The design remains a concept in the mind of the designer until he defines it through verbal description ,sample ,drawing,writing,etc .It defines in advance what the manufacture expects to make .

It defines what the consumer can expect to get . The specification serves as an agreement between manufacture and consumer on the nature of character of the products .

Ø ARTIFICIAL NEURAL NETWORK .

v INTRODUCTION TO ARTIFICIAL NEURAL NETWORK (ANN)

· An artificial neural network or commonly just neural network is a interconnected group of artificial neurons that uses a mathematical or computational model for information processing based on connectionist approval to computation. In most cases an artificial neural network is an adaptive system that changes its structure based on external and internal information that flows through the network .

· In more practical term neural network are non –linear statistical data modeling tools.They can be used to model complex relationships between inputs &outputs to find pattern in data .

INPUT HIDDEN OUTPUT

A neural network is an interconnected group of nodes, akin to the vast network of neurons in human brain.

v BACKGROUND

There is no precise agreed definition as to what artificial neural network is ,but most would agree that it involves, network of simple processing element ( neurons) which can exhibit complex global behaviour , determined by connections between the processing elements & elements parameters .The original inspiration for the technique was from examination of the central nervous system & the neurons (& their axons, dentrites &synapses) which constitute one of its most significant information processing element (see neuroscience).In artificial neural network model, simple nodes [called various“neurons”,”neurodes”,”PEs”(“Processing Elements”) or “units”] are connected together to form a network of nodes. Hence the term artificial neural network . While artificial neural network does not have to be adaptive perse , its practical use comes with algorithm designed to after the strength (weights)of the connections in the network to produce a desired signal flow. These networks are also similar to biological neural network in the sense that functions are performed collectively& in parallel by the units , rather than these being a clear declination of subtasks to which various units are assigned(see also connections). .Correctly the term artificial neural network tends to models designed with emulsion of the central nervous system (CNS) in mind are subject of theoritical neuroscience.

In modern software implementations of artificial neural network the approval inspired by biology has more or less been abandoned for a more practical approach based on statistics & signal processing\n some of these systems neural networks, or part of neural network’s (such as artificial neurons)are used as components in larger systems that combine both adaptive &non –adaptive element . While the more general approach of such adaptive system is more suitable for real world problem solving ,it has far less to do with the traditional artificial intelligence connectionist models .What they do how ever have in common is the principal of non –linear , distributed , parallel and local processing and adaptation .

v MODELS

NN model in artificial intelligence are usually referred to us Artificial neural network(ANN); this are essentially simple mathematical models defining a function f(X)=Y . Each type of ANN model corresponds to a class of such function.

v THE NETWORK IN ANN

The word network in term ”ANN” arises because the function f(x) is defined as a composition of other function gi(x) , which can further be defined as a composition of other function .

This can be conveniently represented as a network structure , with arrows depicting the dependencies between variables . A widely used type of composition is non - linear weighted sum , f(x) = k ∑ wi*gi*(x)

where , k is some predefined function such as hyperbolic tangent . It will be convenient for following to refer a collections of functions gi as simply a vector

g = ( g1 , g2 , -------- , gn )

x h1,h2,h3 g1,g2 f

ANN DEPENDENCIES GRAPH

This figure depicts such a decomposition of fi with dependencies between variables indicated by arrows. This can be interpreted in two ways .

The first view is functional view – the input x is transformed into a 3-D vector h , which is then transformed into a 2-D vector g , which is finally transformed into f . The view is most commonly encountered in context of optimization .

The second view is probablisticsview , the random variable F = f(G) depends upon the random variable G = g(H) , which depends on H = h(X) , which depends on random variable X . The view is most commonly encountered in context of graphical model.

v APPLICATIONS

The utility of ANN lies in the fact that they can be used to infer a function from observation . This is particularly useful in applications , where the complexity of the data or task makes the design of such a function by hand impractical .

v REAL LIFE APPLICATIONS

The task to which ANN are applied tend to fall within the following broad categories :

· Function approximation , or Regression analysis , including time series prediction & modeling .

· Classifuication , including pattern & sequence recognition , novelty detection & sequential decision making .

· Data processing , including filtering , clustering , blind source separation and compression .

Application areas include system identification & control ( vehicle control , process control ) , game playing 7 decision making ( backgammon , chess , racing ) , pattern recognition ( radar system , face identification , object recognition , & more ) , sequence recognition (gestures , speech , hand written text recognition) , medical diagnosis , financial application , data mixing ( or knowledge discovery in data bases , “KDD” , visualization & e-mail spam filtering .

v TYPES OF NEURAL NETWORK

1 . Feed forward neural network..

* Single – layer perceptron .

*Multi – layer perceptron .

*ADALINE.

* Radial basis function .

* Kohanen self – organising network .

2 . Recurrent network .

*Simple recurrent network .

*Hopfield network .

3 Stochastic NN

*Boltzmann machine.

4 . Modular NN.

*Committee of machine.

*Associative NN (ASNN).

v BASIC CONCEPT OF NEURONS WORKING: -

(BIOLOGICAL NEURON)

The human brain is made up of vast network of computing elements, called NEURONS. A neuron is a special cell that conducts an electrical signal. Neurons interact through contacts known as synapes.

A neuron operates by receiving signals from other neurons through synapes. The combination of these signals in excess of a certain threshold or activation level, will result in the neuron firing , i.e.. sending signal on to other neurons connected to it .Some signals act as excitations and other as inhibitors to neuron firing. What we call thinking is believed to be the collective effect of the presence or absence of firing in the pattern of synatic connection between neurons.

At rest, neuron maintains an electrical potential about 40 – 60 millivolts.When fires change in potential to about 90 – 100mV.The impulse travels between 0.5 – 100 mt/sec and lasts about 1 milliseconds .

It is clear that if signal speed or rate were the sole criteria for processing performance, electronic computer would win hands down. What the human brain lacks in these, it makes up in no. of elements and interconnection complexity between these elements.This difference in structure manifests itself in atleast one important way ; the human brain is not as quick as electric computer at arithmetic , but it is many times faster and hugely more capable at recognition of patterns and perception of relationships.

NEURAL NETWORK CONSTRUCTION

Construction of neural network involves following task:-

· Determine the network properties

The network topology (connectivity), the types of connections ,the order of connections, and the weight range.

· Determine the node properties :-

The activation range and the activation (transfer) function.

· Determine the system dynamics:-

The weight initialization scheme, the activation calculating formula, and the learning rule.

Ø LIMITATIONS OF ANN

1. Selecting input is not straightforward .

2. Priortizing cause – specific mortality fraction over sensitivity or specificity is a manual process .

3. Designing optical networks for each cause of death is time – consuming .

4. Sensitivity & specificity may be not high enough for the algorithms to be generalicable to a variety of setting what are limitations of artificial neural network .

An ANN can create a model based only on the data with which it is developed . Therefore , patients with clinical variable that are outside the range of variable used to develop the model will not be able to use this application .

For example : The upper limit of PSA level for the lymph node spread model is 84.2 ng/Ml . This model would be valid for a patient with a pre – treatment PSA that is higher than 84.2 ng/mL . Further this model is designed to predict the risk of lymph node spread in men with cljnically localized prostate cancer only ( not for advanced disease or for men who have not been diagnosed with prostate cancer).

v LIMITATIONS APPLICABLE TO ACTIONS ARISING

FROM DEFICIENCIES IN CONSTRUCTIONS OR IMPROVEMENT IN REAL PROPERTY

No action may be brought to recover damage for injury to property , real or personal , or for an injury to person , arising out of any deficiency in design , planning , supervision or observation of construction ; or construction of an improvement to real property or no action may be brought for contribution or indemnity , against any person , firm or corporation performing , or furnishing the design , planning , supervision of contruction or construction of such improvement to real property more than six years after the written acceptance or actual occupancy or use , whichever occurs first , of such improvement by the owner there of . This limitatjon shall apply to actions against persons , firms and corporation performing or furnishing the design , planning , supervision of construction or construction of such improvement to real property for the state of Mississippi or any agency ,department , institution or political subdivision there of as well as for any private or non govt. entity.

This limitation shall not only appaly to any person ,firm ,or corporation in actual possession and control as owner tenant or otherwise of improvement at the time the defective and un safe condition of such improvement causes injury .

This limitation shall not apply to action for wrongful death .

v LIMITATIONS OF ANN IN CONTEXT OF NUMERICAL WEATHER PREDICTION

Parametric representation , or parameterization is used in numerical modelling of various atmospheric variable . It involves a statistical analysis that enables the representation of true processes by simpler parametric relations . Three purpose may motivate such an analysis :

1 . Getting better understanding of the system .

2 . Allowing a computation of the process that is faster than exact

formulation.

3 . Linking two sets of variable when the exact formulation is not

known.

Numerical application in numerical weather prediction fall in one of this category . This is the case of all the components of an atmospheric and / or oceanic forecast model , of the retrieval of geophysical parameters from satellite data , & of the analysis of satellite imagely .

Ann’s like the multi – layer perceptron provide powerful solution for this problems , but have been only marginally used in operational weather centers . The novelty of the approach partly explains this situation . Several reasons can be found in some difficulties associated to the ANN approach as well . Among others , we may mention the remote link of a trained ANN with the physical laws , the difficulty of gathering an adequate (i.e.. representative and correctly sampled ) training date set for high dimension problem , the difficulty of having both the direct model and its derivatives right ;or the difficulty of modifying the ANN once it has been trained.

v PERCEPTRON FUNCTIONAL LIMITATION OF TWO LAYER

It is that it can only recognize linearly seperable patterns due to only having one adaptive layer .A lineary seperable pattern is one that can be separated into two distinct classes by drawing a single line .

v CEMENT MANUFACTURING PROCESS IN ACC

· CRUSHER DEPARTMENT :-

In the crushing section limestone coming from mines is crushed to a preferable & required size. In crushing two types of crushers used:-

(i) Primary Crusher with capacity

(ii) Secondary Crusher with capacity

· LIMESTONE SIZE:-

About 5kg sample of crushed stone is drawn, daily from alternate crusher (if both crushers are running), and tested for size distribution on different sieve sizes, starting from ¾ inch to 1/8 inch. It gives indication of performance of secondary crusher hammers; condition of screen used for returning back over-sized crushed stone to crusher for recrushing. Results are communicated to crusher in charge by format QC/DUP/CR for action in respect of replacement of crusher hammers, repairing/replacing the screen.

· ACCEPTABLE SIZE:

Minimum 95% of limestone sample, analyzed for sieve analysis, should pass ¾” sieve

· WORKING PROCEDURE:-

Limestone from the mines is carried or brought to crushing section with the help of dumpers. Limestone is fed to feeder & then to a primary crusher, which is a Jaw crusher with capacity ( ton ) The crusher stone then goes to vibrofeeder & then limestone belt conveyor.

The crushed limestone, then are fed to secondary hopper distributed limestone to secondary crusher with capacity ( ton). Here the limestone is crushed to very small size. The product from the crusher fed to another vibrating screen the limestone of size 12.5 mm. Then the crushed limestone is brought up to gantry. From gantry the limestone is passed to raw mill section. The crushed limestone is stored in the gantry. It is carried to the gantry with means of the crane. With the help of the crane, the crushed limestone is fed to the hopper of the raw mill.

Limestone is supplied from patharia and jamul mines to works. The limestone size is reduced to about ¾” size or less in crusher, stored in gantry and then sent to raw mills for pulverizing. Depending upon the quality requirement the same are crushed together in set proportions or independently and stored in limestone gantry accordingly i.e. to separate heaps for high and low grade stones are made in crushed limestone gantry. If total carbonates (T.C.) of high grade limestone is observed lower than norms, the limestone is taken in low grade limestone chute. If low grade limestone T.C. is lower than norms or MgCO₃ of limestone is more than the norms then it is stored

v RAW MILL (Pulverizing of limestone)

The crushed limestone powder is fed into limestone hopper. The limestone hopper passes to table feeder & then to belt conveyor. This belt conveyor takes the powered limestone to bucket elevator. Bucket elevator is vertical column contain no of buckets moving upward. These buckets carry away the powered limestone & feed them into feed screw. Feed screw causes the fine powder to move forward in horizontal direction. This screw brings the powered limestone into the air separator column. Where the blower blows air, the air causes very fine & light particle to move out in the outer chamber & the heavier & coarser particle still sticks with feed blower and passes inside inner chamber. From the inner chamber these coarser particle are feed into rotating raw mill. Raw mill is horizontal tube cylindrical in shape. Inside this tube spherical balls are placed, when the mill rotates cylindrical balls also tends to rotate & the limestone particle coming in between is grinded into very fine particle. Three closed circuit raw mills are provided for grinding of limestone. Limestone is fed through over head crane to raw mill hoppers on the instruction from the laboratory through lab Asstt /Tester. Each mill is provided with feed tables for proportioning of required dosage of limestone to give desired raw meal quality.

For better grinding, proper mill ventilation is recommended which is checked by manometer as draught through mill. If any time residue is observed above norms then corrective action is taken jointly by tester/miller so that it comes in range.

· RAW MEAL RESIDUE ON:

Rawmeal/KF Rawmeal/KF

LEPOL HUMBOLDT

212 micron 2.5% max 7.0% max

90 micron 18.0% max 28.0% max

· RAW MEAL:-

Quality of raw meal (the product of raw mill) is controlled by drawing and testing samples from F.K. Pumps/ D Pumps of running raw mills. All samples are tested for T.C. (total carbonates) and alternate samples are tested for residue on 90 and 212 microns. The samples are drawn through sampler at almost fixed time. Total carbonate deviations are controlled by making adjustment in feed table settings or mixing and feeding of low and high grade limestone in hopper in required proportions. Residues are controlled by Tester / Lab Asstt. along with miller and result of the same is communicated to mills.

v KILN DEPARTMENT:-

At the kilns the process of clinker formation takes place. There are three kilns at the ACC, no. 1&2 operate with the semi dry process while no. 3 kiln operates with the dry process. At the kilns, the kiln feed is charged at one end and coal firing is done at the other end.

· The kiln feed—

q Raw material:-

The basic ingredient for Portland cement consists of limestone, sea shells, meal or chalk that provides the iron components. The no of raw mill required at any one plant depend upon the composition of this material & type of cement being produced. To affect the proper, raw material are continually sampled & analyzed & proportions adjusted as they are together.

q Coal:

Coal is used as fuel for firing in kilns and in slag dryers. Coal is received from collieries as per linkage allocated by ministry of India .Coal is also received from other sources as per work order, copy of which maintained at q.c. and purchase deptt. Coal is received by rail as well by as road.

q Acceptable Norms:

-High grade: Total Carbonate (T.C.) 79.5%

& above

Magnesium Carbonate max 5%

-Low Grade: Total Carbonate 70% to79.4%

Magnesium Carbonate max 5%

After being excavated in the quarry mine, limestone is first passed through primary crusher then to secondary. In modern plant this sampling & testing is the source of data fed into a digital computer controlling composition of stored & blended raw fed.

· KILN:-

At the Jamul Works Clinker is produced by two different types of processes

1. Dry process (Humboldt kiln)

2. Semi dry process (Lepol kiln)

Raw mill thus ground is stored separately suiting to the requirement for two types of processes. Raw meal total carbonates target for Humboldt kiln and Lepol kiln is different due to difference in process.

In Lepol kiln raw meal is converted into nodules in noduliser by feeding raw meal and spraying water in it. These nodules are fed to Lepol kiln from inlet end and pulverized coal is fired from outlet end and thus clinker is formed. It is same as ball mill used for dry grinding.

In Humboldt kiln fine powder of raw meal is fed to cyclone from where it enter to kiln and pulverized coal is fired from outlet end and thus clinker is formed

· FINE COAL: (for process efficiency only)

Residue on 90 micron:

For kiln 22% max

For MFC 38% max

· Moisture:

During monsoon (jun-oct) - 7% max

During normal season - 5% max

· NODULISER:-

Noduliser is used for producing nodules in semi dry process of clinker formation .Where inclined disc rotates with spray of water & form nodules.

For the Lepol kilns nodules made in the noduliser are shift wise tested for moisture sieve analysis and also total carbonates. Noduliser operator adjusts its operation to produce within stipulated norms. The nodules characteristics are important from process efficiency point of view only.

Lepol nodules moisture 10% - 12%,Litre Weight 1050 – 1200 gm/liter and size +15 mm 25% max and 10 mm 35% max.

· Nodules & Calcinations:-

MAX MIN

Porosity of Lepol nodules 26% 19%

Green strength of Lepol

nodules 4.0 kg 1.0kg

Humboldt kiln (%)

Preheater (cyclone 1) 15.0 7.0

Preheater (cyclone 2) 26.0 18.0

Preheater (cyclone 3) 42.0 30.0

Preheater(cyclone4) 85.0 65.0 NMFC 65.0 50.0

· LEPOL GRATE:-

After producing nodules by noduliser, it fed into lepol grate .Where moisture is removed and initial calcinations take place.

· BURNING:-

All kiln are provided with one coal mill each, for fine coal preparation , fired in kiln for pyroprocessing , along with to more coal mills, one as stand by and other for feeding fine coal for MFC firing . Usually kilns are fired with fine coal from their own coal mill except during initial light up or during coal mill break down , when fine coal is taken from coal mill no 4 or no 5 ,which are tuned for coarser coal grinding ,for MFC and slag drier in which fine coal is fired at higher residue.

Clinker produced is transported by belt conveyor to storage gantry.

· CHARGING AND DRYING:-

Available raw material to produce a cement that meets the specialized composition, the feed is now ready to be introduced into the kiln within the kiln the process of making the chemist clinker is accomplished in four zones; Dehydration, Calcinations, Clinker zing & cooling .The reactions taking place in each of the zones are important steps toward production of high quality clinker. None of these can be neglected or passed by if kiln is to operated properly, experience has proved that a kiln can never be properly operated without consideration of all the reaction start taking place as shown in the table:-

· REACTION TEMPERATURE WITHIN THE KILN:-

Approximate Temperature ( 0C )	Reaction
100	Evaporated of free water from the feed
500	Evolution of combine water from clay
750	Evolution of CO2 from limestone. Start of calcinations.
800 to 900	Formation of dicalcium silicate
1095 to 1205	Formation of tricalcium aluminate & tricalcium aluminoferrite (C3AF)
1260 to 1455	Formation of tricalcium silicate with progressive Disappearance of free lime (CaO)

· CALCINATION:-

In the Jamul Cement Works cement manufacture is done by dry & semi dry process both. In the dry process feed is in the form of dry power. In this factory, there is a preheater arranged. It consists of FIVE CYCLONE in series through which limestone passes. There are three Kilns in this plant, about 94 % calcinations of limestone power is composed in these preheater before they enters the kiln. Rest is done in kiln.

Calcinations is the process through which CO₂ is driven off the limestone, leaving lime.

CaCO₃ ------------ CaO + CO₂

At a temperature of approximate 750 ⁰C calcinations begins. Carbon dioxide disassociates from the feed & is carried away by the kiln gases. Kilns feed that is not completely burned & is one of the kiln feed before it enters burning zone is essential to proper burning of the clinker

· CLINKERIZATION:-

In transition zone C₃A, C₃AF, & C₂S are formed. In the burning zone the lime rich mixture containing silica, alumina & ferric oxide, with small percentages of other, is heated up to the sintering temperature & becomes viscous .This is called clinkerization & is manifested by the liquefying of the clinker constituents. As pointed out earlier, formation of the most important compound C₃S, does not start until the feed has reached a tem of approximately 1260 C. Between 1350 to 1400 C the free lime has completely disappeared. This is the tem. To which the clinker has to be burned in ordered that it can be good quality.

Clinker normal kiln feed and clinker chemical composition (in %) of clinker produced are given below.

	KILN FEED	KILN FEED	CLINKER
	LEPOL	HUMBOLDT	(AVERAGE)
SiO2 %	11.5-14.5	10.5-12.5	20.5-22.5
Al2O3 %	3.0-4.0	2.8-4.0	5.5-7.0
Fe2O3 %	1.8-2.8	1.8-2.8	2.8-4.4
CaO %	42.5-43.9	43.0-44.6	MIN. 62.5
MgO %	MAX 2.6	MAX 2.6	MAX 3.5
LiO %	34.0-35.7	34.5-36.5	MAX 1.5
MgCO3 % Free Lime	5.0 MAX	5.0 MAX	----- Max. 3.5 %
SO3 %			MAX 0.5%

· OPERATING RANGES OF RAW MEAL/ KILN FEED T.C

q LEPOL KILN TOTAL CARBONATES: Usually 77.5% to 79.5% compatible with coal ash. If total Carbonate (T.C.) of high grade limestone is observed lower than norms, but within low grade limestone norms, the limestone is taken in low grade limestone chute .If low grade limestone T.C. is lower than norms or MgCO₃ OF high or low grade limestone is more than the norms then it is stored separately in crushed limestone gantry. Mines are communicated of the analysis results and advised to take corrective action at the loading point .The separately stored limestone is used in small proportion with acceptable quality of limestone.

q HUMBOLDT KILN TOTAL CARBONATES:

Usually 79.5% to 81.5% compatible with ash. However, if CaO in clinker is not up to desired level, this range will be changed, the ultimate aim to get required clinker CaO. Kiln feed samples are drawn thrice in a shift from Humboldt kiln and twice in a shift from Lepol kiln at almost fixed time. All samples are tested for total carbonates and residue on 212 and 90 microns sieves. MgCO₃ is tested once in shift. If any deviations are observed than norms KF silos are changed, ultimately aim is to get required clinker quality. The quality parameter and ranges for KF will be same as meal.

· COOLING:-

The process when the material has passed, the hottest area of the burning zone, as the manner in which the clinker is cooled affects its quality to a great extent. Once formation of the compound has been completed & the clinker has reached at temperature of approximate 1370 ⁰C, the process of cooling starts. Cooling of clinker start from few feet short of the discharge end of the kiln in what is considered still part of the burning zone. The speed of the cooling required until all liquid in the clinker solidified is important. Rapid cooling is beneficial for quality of the clinker & grind ability.

v CEMENT MILL:-

At the cement mill clinker, slag and gypsum are mixed in proportion and grinded to get cement.

q Gypsum :

Gypsum is used at cement grinding stage on control setting of cement. Jamul cement works receives mineral gypsum, marine and chemical anhydrite gypsum from various sources by truck and rail. Works order are issued by Jamul cement works /material management division (mmd) Mumbai .copies of work order are maintained by q.c. and purchase deptt

q Slag:

Slag is the major component in Portland slag cement and is used at cement grinding stage .It is procured from Bhilai Steel Plant as granulated blast furnace slag . the slag is procured as per contract from (BSP) by road transportation .The slag is having moisture usually above 10% and is dried in slag drier feeding it to cement mill.

q Cement Grinding:

Jamul works had got 8 open circuit cement mill .Since Portland slag cement (psc) is only produced , clinker ,slag, and gypsum in set proportion are ground in cement mill to form cement. Clinker, slag and gypsum are separately filled up in mills hoppers and fed to mill .in some of the mills gypsum is also proportioned with clinker in gantry itself and fed to mills through clinker hoppers. The cement ground is stored in cement storage silos.

q THE CEMENT HAS THE FOLLOWING CONTENT:

CONTENT	PERCENTAGE
SiO2	22% - 28%
Al2O3	11% - 15%
Fe2O3	1.5% - 2.5%
CaO	42% - 50%
Chloride	0.05% max
Sulphide Sulphur	0.8% max

v THE NORMS FOR QUALITY OF CEMENT GRINDING, TESTED AT SITE LAB, WHICH IS IN AMBIANT CONDITION.

SPECIFIC SURFACE:

320 – 380 m²/kg

RESIDUE ON:

90 micron: 6.0 % maximum.

SETTING TIME (IN MINUTES)

Initial : 60-180

Final : 90-220

SO₃ :1.5 -2.5 %

CaO : 42 – 50%

· CEMENT TEMPERATURE:

130 deg C Maximum

Clinker Temp (Feed table) : 75 deg C Minimum

· GYPSUM:

Moisture: 5 % max (during normal season)

: 10 % max (during monsoon)

Silica modulus (S.M.) 1.9-2.5

Alumina modulus (A.M.) 1.5-2.1

Lime saturation factor (LSF) 0.86-0.97

Clinker produced beyond norms is segregated separately at earmarked place and later on used in small proportions with good quality clinker.

· CLINKER:

Liter weight 1200-1400 gm/liter

CaO 62.5% minimum

q DRIED SLAG:

Moisture in dry slag is monitored and corrective action at the dryer end is taken if the moisture exceeds the control limit.

Gypsum is also tested for moisture content. If gypsum moisture content is more then gypsum is fed from other spot in gantry to mill hopper.

Clinker and slag proportioning is done by adjusting the feed rate of the same, to achieve the targeted cement composition for CaO, SO₃ and specific surface

After grinding in the cement mills the cement is stored in the cement silos from where it is conveyed to the packing department for packing.

Ø INTRODUCTION TO MATLAB :-

v ABOUT MATLAB.

MATLAB is a software package for high-performance numerical computation and visualization. It provides an interactive environment with hundreds of built-in functions for technical computation, graphics, and animations. Best of all, it also provides easy extensibility with its own high-level programming language. The name MATLAB stands for Matrix Laboratory.

MAT Lab’s built in functions provides excellent tools for linear algebra computations, data analysis, signal processing, optimization, numerical solutions of ordinary differential equations (Ode’s). There are numerous functions for 2-D and 3-D graphics as well as animation. Also, for those who cannot do without their FORTRAN or C-codes, MATLAB even provides an external interface to run those programs from within MATLAB.

The basic building block of MATLAB is the matrix. The fundamental data-type is the array. Vectors, scalars real matrices and complex matrices are all automatically handled as special case of the basic data-type. MATLAB loves matrices and matrix operations. The built in functions are optimized for vector operation. Consequently, vectorized commands or codes run much faster in MATLAB.

There are also several optional Toolboxes available from the developers of MATLAB.

These Toolboxes are collections of functions written for special applications such as Symbolic Computation, Image Processing, Statistics, Control System design, Neural Networks, etc.

· BASICS OF MATLAB :

* MATLAB Windows:

On almost all systems, MATLAB works through three basic windows, which are discussed below.

1. Command Window: This system is the main window. It is characterized by the MATLAB command prompt ’>>’. In MATLAB 7 this window have four sub windows they are:

· Launch Pad: This sub window lists all MATLAB related applications and toolboxes that are uninstalled on your machine. We can launch any of the listed applications by double clicking on them.

· Work Space: This sub window lists all the variables that we have generated so far and shows their type and size.

· Command History: All commands typed on the MATLAB prompt in the command window get recorded, even across multiple sessions.

· Current Directory: This is where all our files from the current directory are listed. We can do file navigation here.

2. Graphics Windows: The output of all graphics commands typed in the command window are flushed to the graphics or figure window, a separate gray window with white background color.

3. Edit Windows: This is where we write, edit, create, and save our programs in files called ‘M- files’.

vNEURAL NETWORK TOOL BOX IN MATLAB:-

MATLAB is a very powerful tool for calculation, visualization and programming. In addition to the pure mathematical part of MATLAB there are several toolboxes available to expand the capabilities of MATLAB; the Neural network Toolbox (NN Toolbox) is one of these toolboxes. The neural network toolbox makes it easier to use neural networks in MATLAB. The toolbox consists of a set of functions and structures that handle neural networks. This is good because we do not want to write code for all activation functions, training algorithms, etc. the elements of MATLAB and the neural network toolbox are more easily understood when explained by example.

Following is introduction, how to use neural network commands in command box.

Using these procedure and commands program of neural network can be written and executed. Various algorithms, functions and commands are used to write a program in MATLAB for implementing neural network.

v The structure of Neural Network Toolbox: -

The toolbox is based on the network object. This object contains information about everything that concern the neural network.

It contains the training, initialization and performance functions.

The trainFcn and adaptFcn are essentially the same but trainFcn will be used in this tutorial. By setting the trainFcn parameter you tell MATLAB which training algorithm it should use. The ANN toolbox includes almost 20 training functions. The performance function is the function that determines how well the ANN is doing its task. The initFcn is the function that initializes the weights and biases the network. To get a list of the functions those are available, type help nnet. To change one these functions to another one in the toolbox or one that you have created, just assign the name of the function to the parameters eg.

net.trainFcn=’mytrainingfun’

To create a new feed forward neural network uses the command newff. You have to enter the max and min of the input values, the number of neurons in each layer and optionally the activation function.

>>net = newff([ 0 1;0 1],[2 1],{‘logsig’,’logsig’})

The variable net will now contain an untrained feed forward neural network with two neurons in the input layer, two neurons in the hidden layer and one output neuron, exactly as we want it. The [0 1;0 1] tells MATLAB that we want to use logsig function as activation in all layers. The first parameter tells the network how many nodes there should be in the input layer; hence you do not have to specify this in the second parameter. You have to specify at least as many transfer functions as there are layers, not counting the input layer. If you do not specify any transfer function MATLAB will use the default settings.

The command, sim is used to simulate the network and calculate the outputs, for more info on how to use the command type helpwin sim . the simplest way to use it is to enter the name of the neural network and input matrix, it returns an output matrix.

· Import Data: -

The file is loaded into MATLAB with the command load bc-repl.data. the contents is automatically placed in the matrix bc-repl. Note that MATLAB may not show the data as it was shown in the data file. MATLAB automatically chooses a factor that makes the matrix look better. The data loaded is the same as as in the data file, its just that MATLAB does not show it in the same way. The first column contains the ID’s of the patients, and since we are not interested in this we construct a new matrix without it.

· Training algorithms: -

In the neural network toolbox there are several training algorithms already implemented. That is good because they can do the heavy work of training much smoother and faster than we do by manually adjusting the weights. Now let us apply the default training algorithm to our network. The MATLAB command to use is train; it takes the network, the input matrix and the target matrix as input. The train command returns a new trained network. For more information type help win train. If the gradient of performance is less than .min_grad the training is ended. The time component determines the maximum time to train.

Activation Function: -

When a neuron updates it passes the sum of the incoming signals through an activation function, or transfer function as MATLAB calls it. There are different types of activation functions, some are saturated and assures that the output value lies within a specific range, like logsig, transig, hardlims and satlin. Some transfer functions are not saturated like purelin. Some of the transfer functions in the neural network toolbox are plotted in the figure below.

The transfer function is chosen when you create the network, and is assigned to each layer. To create a feed forward network with a layer of two input neurons, three transig neurons in the hidden layer and one logsig neuron in the output layer enter:

>>net = newff([0 1;0 1],[3 1],{‘transig’,’transig’});

BLAST FURNACEè

Blast furnace is an equipment where various raw materials (iron ore,limestone as flux, quartzite as additives ,coke as fuel,sinter as Fe-bearing materials,Mn ore) are being heated at high temperature to convert it into

Pig iron.

IRON MAKING IN BLAST FURNACES:

TheBlast furnaces department consist of seven blast furnaces. Three of them are of 1033 cu.m. useful volume (1200T/day) three of 1719 cu.m. useful volume (1800T/day) and BF-7 is of 2002 cu.m. volume (2658T/day).The annual rated capacity of hot metal is 4.25 MT with 350 working days in a year .

THERE ARE THREE MAIN SECTIONS OF THE DEPARTMENT:

1. RAW MATERIALS SECTION

2. FURNACE SECTION

3. AUXILIARY SECTION

Slag processing plant and G.C.P. although do not come under B.F. department , are integrated units. An express lab is provided to ensure rapid analysis and quality control .

1.RAW MATERIAL SECTION:

The raw materials used for the production of pig iron are:

a. Coke (fuel )

b. Iron ore

c. Sinter(Fe – bearing materials)

d. Lime stone(flux)

e. Mn ore

f. Quartzite (additives)

Ore yard

Raw materials arriving in blast furnace department from various sources are unloaded in the Ore trench of ore yard.

Ore yard is located towards blast furnace. The ore yard is meant for stocking and averaging materials .Approximately, the following amount of materials can be stored in ore yard at a particular time.

1. Fine ore : 85,000 tonnes

2. Iron ore (lumps): 1,80,000 tonnes

3. Lime stone : 40,000 tonnes

4. Mn ore : 11,000 tonnes

5. Quartzite : 3,000 tonnes

Raw materials from ore yard are charged by means of ore bridge cranes (obc) into 65 tonnes electrically operated transfer cars . They carry the materials into the respective bunkers.

STOCK HOUSE:

In the conventional system all raw materials from the bunkers except coke were drawn by scale car which transport the materials the loads into the skips . The scale cars operator was required to maintain the sequence of drawing materials from the different bunkers which helped in averaging the chemical composition .

Under modernisation programme the bunker gates have been modified and the stock house has been completely conveyorised .

Bunkers->Vibrofeeder->conveyor belts->weighhopper->skip car

COKE CHARGING SYSTEM:

Coke from coke bunkers falls on a vibratory screen BF size coke from the screen collects in a weighing hopper.

HOIST HOUSE:

For taking charged materials to the furnace top , two way skip hoist with two skips of useful volume are provided. Bell hoist , equalizing valves , test rods ,etc are operated from hoist house.

2.FURNACE SECTION:

Blast furnace is a vertical shaft furnace , enclosed in a welded shell , limed with fire- clay bricks of high alumina content . The hearth bottom, hearth,bosh and cylindrical part of the bosh are cooled by means of plate cooler.

a. Top charging equipment.

b. Charging programme

c. Stoves

d. Flushing of slag , tapping and disposal of hot metal

3.AUXILIARY SECTION:

The auxiliary section of blast furnace consist of the following section:

a. Ladle repair shop

b. Pig casting machine

c. Cold pig yard

d. Taphole mass shop

e. Slag dump post

f. Slag ladle depot

SPECIFICATION OF RAW MATERIALS:

MATERIAL	CHEMICAL ANALYSIS	SIZE
Iron ore	Fe SiO2 P Al2O3	64+1.5% 10 to 50mm 1.8% 0.10% Max 1.5%
Lime stone	CaO SiO2 MgO	42+1% 25 to 50mm 6.5% 8%
Mn ore	Mn SiO2 Al2O3 P	30-35% 25to 80mm 20-25% 5% 0.3%Max

CONTROLLABLE PARAMETERS OF BLAST FURNACE:

Operation parameters
`	Blast time	Temperature
atm	hrs	C
0.8 to 1.5	~2	950-1000
1	2	3
0.89	1.99	987
0.94	1.89	988
1.38	1.92	975
1.49	1.95	977
1.48	1.97	967
1.44	1.99	999
1.46	1.89	980
1.49	1.88	986
1.32	1.87	957
1.31	2.03	987
1.22	1.96	988
1.24	1.99	998
1.34	1.88	992
1.21	1.98	992
1.21	2.00	999
0.88	2.01	987
0.87	1.99	968
0.86	1.89	966
0.89	1.98	965
0.92	1.97	964
0.99	1.99	976
1.43	1.89	999
1.22	1.88	985
1.28	1.95	980
1.24	1.96	981
1.25	1.99	980
1.28	1.87	980
1.00	1.98	980
1.06	1.98	980

UNCONTROLLABLE PARAMETERS OF BLAST FURNACE

Coke Analysis
Blast Furnace Ref	Ash	VM	M	S	C
No.	%	%	%	%	%
1 to 7	15 to 16	0.3 to 0.4	5+/-0.5	0.5 to 0.6	75 to 80
22	23	24	25	26	27
1	15.01	0.30	5.06	0.53	75.77
2	15.05	0.32	5.01	0.50	78.65
3	15.60	0.31	5.00	0.55	76.00
4	15.07	0.34	4.97	0.59	75.01
6	15.22	0.31	4.44	0.53	78.09
7	15.69	0.35	4.54	0.50	78.22
1	15.55	0.33	4.52	0.51	75.00
2	15.54	0.30	4.68	0.56	76.07
3	15.14	0.33	5.04	0.58	75.05
4	15.22	0.36	5.06	0.59	77.00
6	15.24	0.37	4.74	0.58	75.04
7	15.40	0.38	4.73	0.51	78.55
1	15.41	0.37	4.71	0.55	76.67
2	15.43	0.37	4.70	0.50	76.77
3	15.44	0.38	4.68	0.55	76.57
4	15.46	0.30	4.66	0.55	76.57
6	15.47	0.32	4.65	0.55	76.57
7	15.49	0.31	4.63	0.55	76.57
5	15.51	0.34	4.62	0.55	76.58
5	15.52	0.31	4.60	0.55	76.58
5	15.54	0.35	4.58	0.55	76.58
5	15.55	0.35	4.57	0.55	76.58
6	15.57	0.35	4.55	0.55	76.58
6	15.58	0.35	4.54	0.55	76.58
6	15.60	0.35	4.52	0.55	75.09
6	15.62	0.35	4.50	0.55	75.97
1	15.63	0.35	5.06	0.56	76.00
4	15.65	0.35	5.04	0.56	76.22
5	15.66	0.35	5.33	0.56	76.20

UNCONTROLLABLE PARAMETERS OF BLAST FURNACE:

Blast Furnace Gas				Blast Furnace Slag
CO	CO2	N2	H2	SiO2	Al2O2	CaO	MgO	MnO	CaO/SiO2
%	%	%	%	%	%	%	%	%	Basicity
25 TO 26	15 TO 16	55 TO 57	2 TO 3	32 to 33	19 to 20	31 to 33	9 to 10	<>	0.98
28	29	30	31	32	33	34	35	36	37
25.07	15.52	55.04	2.06	32.07	19.88	31.07	9.00	0.56	0.98
25.99	15.54	55.89	2.09	32.65	19.86	31.66	9.80	0.76	0.98
25.28	15.55	55.86	2.67	32.78	19.56	31.56	9.55	0.65	0.98
25.67	15.57	56.00	2.56	32.55	19.67	31.76	9.56	0.77	0.98
25.89	15.58	56.00	2.58	32.54	19.88	31.77	9.57	0.66	0.98
25.43	15.60	56.22	2.65	32.28	19.67	31.89	9.67	0.76	0.98
25.40	15.62	56.78	2.79	32.45	19.50	31.54	9.55	0.78	0.98
25.61	15.63	56.05	2.78	32.45	19.14	31.66	9.54	0.80	0.98
25.05	15.65	55.84	2.67	32.55	19.34	32.47	9.66	0.82	0.98
25.06	15.66	55.88	2.66	32.28	19.67	32.44	9.71	0.85	0.98
25.74	15.68	55.67	2.67	32.56	19.79	32.41	9.74	0.87	0.98
25.55	15.69	55.55	2.51	32.66	19.45	32.40	9.77	0.21	0.98
25.66	15.71	55.06	2.15	32.54	19.42	32.22	9.81	0.33	0.98
25.67	15.72	55.09	2.41	32.50	19.39	32.45	9.84	0.34	0.98
25.44	15.74	55.87	2.44	32.45	19.36	32.78	9.87	0.54	0.98
25.01	15.76	55.05	2.43	32.33	19.33	32.50	9.67	0.87	0.98
25.39	15.77	55.23	2.53	32.53	19.30	32.34	9.57	0.76	0.98
25.38	15.79	55.54	2.66	32.55	19.27	32.32	9.88	0.67	0.98
25.37	15.80	56.00	2.76	32.65	19.24	32.17	9.85	0.66	0.98
25.36	15.82	55.44	2.88	32.66	19.21	31.18	9.45	0.76	0.98
25.35	15.83	55.41	2.56	32.75	19.18	31.84	9.65	0.78	0.98
25.35	15.85	55.38	2.15	32.57	19.65	31.54	9.66	0.89	0.98
25.34	15.87	55.36	2.18	32.77	19.66	31.79	9.54	0.34	0.98
25.33	15.88	55.33	2.52	32.74	19.88	32.20	9.54	0.46	0.98
25.32	15.90	55.30	2.52	32.60	19.87	32.22	9.65	0.49	0.98
25.31	15.02	55.27	2.52	32.65	19.85	32.24	9.54	0.58	0.98
25.30	15.00	55.25	2.52	32.62	19.44	32.26	9.11	0.89	0.98
25.29	15.06	55.22	2.52	32.80	19.65	32.28	9.29	0.68	0.98
25.28	15.88	55.19	2.52	32.73	19.70	32.29	9.00	0.66	0.98