Air Cooled Blast Furnace Slag & Natural Geological Materials

(1)

MASTER’S THESIS

2008:0105 PB

Zana M. Sherif

Analysis of data obtained from Percostations

Differences between data obtained from

Air Cooled Blast Furnace Slag & Natural Geological Materials

M.Sc. in Environmental Engineering CONTINUATION COURSES

Luleå University of Technology

Department of Civil and Environmental Engineering Division of Soil Mechanics and Foundation Engineering

(2)

Analysis of data obtained from Percostations Differences between data obtained from

Air Cooled Blast Furnace Slag & Natural Geological Materials

Zana M. Sherif

(3)

To my Family

Especially My Father

(4)

Foreword

This thesis is a part of the Master Program in Environmental Engineering at Luleå University of Technology, Department of Civil, Mining and Environmental Engineering. The work was performed in cooperation with the Swedish Road Administration Northern region.

Thank you Professor Sven Knutsson my supervisor at Luleå University of Technology, Johan Ullberg my supervisor at the Swedish Road Administration Northern region (Vägverket). There was also great support from Staffan Rutqvist in SSAB- Luleå. Finally my special thanks to my family and all the people who supported me, I couldn’t do this work without you all.

Luleå November, 2008 Zana M. Sherif

(5)

Abstract

During rehabilitation of 16 km at the road 563 between Sjulsmark and Ersnäs Norrbotten County, Sweden in 2006, a 800m long test road was constructed. Three different fractions of Air Cooled Blast Furnace Slag were used. The fractions were 0/63, 0/125 and unsorted material. In addition normal crushed stone was used in one section as a reference. The test road was built by SSAB Tunnplåt AB, SSAB Merox, the Swedish Road Administration Northern region and BDX AB. The purpose was to evaluate Air Cooled Blast Furnace Slag as a road construction material and compare it with normal geological rock materials. A second purpose was to update “ATB Väg” and “ATB Hyttsten”. The test road was divided into 8 parts each part being 100m long. This study deals with the fifth part of the test road, which was constructed using unsorted Air Cooled Blast Furnace Slag and compacted by using ring compactors.

Air Cooled Blast Furnace slag (Hyttsten) is a byproduct produced in steel industry was used for 30 years.

A Percostation, which is used to measure and visualize measured parameters like dielectric constant (Dielectricity), Electrical Conductivity and Temperature, was installed in the studied part to collect data from 5 different depths (15cm, 30cm, 50cm, 80cm and 110cm) in the road section. It is normally used in detecting critical bearing capacity of road materials and to assess the frost and water effects on the road materials.

Data obtained from the percostation in the test road was analyzed and compared with data obtained from two other stations in Finland. The purpose was to investigate the differences between the behaviors of Air Cooled Blast Furnace Slag, which was used to construct the test road, with the behavior of natural geological road materials, which were used in the two Finnish stations. The idea was to be able to explain the facts that may cause the difference between Air Cooled Blast Furnace Slag and other road construction materials.

This study showed that Air Cooled Blast Furnace Slag due to its low water content will have lower Dielectricity values compared with the other road construction materials being used in the other test stations. And since Air Cooled Blast Furnace Slag contains a lot of Sulfur and Vanadium, that may leach down to lower layers or may help to produce H₂SO₄. The Electrical conductivity values in the layers beneath Air Cooled Blast Furnace Slag were found to be will be higher than those measured at the other test sections.

(6)

1. Introduction

1.1. Background

Air Cooled Blast Furnace slag (Hyttsten), which is a byproduct produced in steel industry was used since 1970’s to replace natural materials in road construction. Although the material was used for 30 years only few documents were found in these regards and most of them were in Swedish.

Both the Swedish Road Administration (Vägverket) and the Swedish Environment department (Miljödepartment) aim to reduce using of the natural road construction materials through recycling and reusing materials without affecting the environment.

In August- September 2006 a test road of 800m length was constructed within the road 563 between Sjulsmark and Ersnäs. This project was done in cooperation between the Swedish Road Administration Northern region (Vägverket), SSA Tunnplåt, SSAB Merox and BDX AB. The project aimed to study the properties of Air Cooled Blast Furnace Slag used in road construction and compare it with the natural road construction materials, and to update the “Common Technical Description for Road Building”- “ATB Hyttsten”.

A Percostation System, which is a device to measure and record dielectric constant (Dielectricity), Electrical Conductivity and Temperature that are used to assess the frost and water effects on the road materials was installed under a part of 100m length under the road where Air Cooled Blast Furnace Slag size 0/125 and compacted by ring rollers was used. The Percometers- sensors of the Percostation- were installed at 15cm, 30cm, 50cm, 80cm and 110cm depth.

In this study the data collected from a Percostation installed under the test road that is called Luleå Percostation in this report, was analyzed and compared with data collected from two Percostations in Finland, which are Inari and Koskenkylä. SPSS was used to perform statistical analyses also Microsoft Excel was used to perform further analyzes

1.2. Objectives

The objective of this thesis is to analyze the data collected from Luleå Percostation, where Air Cooled Blast Furnace Slag was used and compare it with the data collected from Inari and Koskekylä Percostations in Finland, where the construction material was considered to be natural materials.

Analyses were performed to answer the following questions:

 What are the differences between the variables Dielctricity, Electrical Conductivity and temperature of Air Cooled Blast Furnace Slag, which was used in Luleå and the natural road construction materials, which was used in Inari and Koskenkylä?

 Why are the variables Dielctricity, Electrical Conductivity different in the three sites?

1.3. Methods

To get the results and conclusions mentioned in this report, the following methods were used:

 A literature study was performed to study Air Cooled Blast Furnace Slag and its chemical, physical and mechanical properties. Also a study was performed to understand the Percostations and the way they work and perform.

 Interviews and internet communications with Vägverket, SSAB and Roadscanners experts were performed to find references and to get information on the test road and general information on Air Cooled Blast Furnace Slag and Percostation techniques.

 Percostations data were obtained in Microsoft word and excel from Roadscannares. More data in Microsoft excel format were obtained from Swedish Meteorological and Hydrological Institute- Environment & Safety Services on Temperature, Precipitation and Relative humidity.

(10)

 Analyses on the collected data were performed by using SPSS and Microsoft Excel to investigate the data quality and interrelationships between the variables.

1.4. Contents

This report contains the following chapters:

 Chapter 1 is an introduction into this master thesis as background, objective, method.

 Chapter 2 describes Air Cooled Blast Furnace Slag based on the literature study. First it gives general information on production process then it explains Air Cooled Slag’s chemical composition and its chemical, physical and mechanical properties.

 Chapter 3 deals with Percostations and there components and the way they work and collect data.

 Chapter 4 describes the test road and its structure.

 Data analyses start in chapter 5 by assessing the quality of collected data. Then the results are explored in chapter 6 and discussed in chapter 7. The details are in tables, figures and diagrams attached as Appendixes 1 to 8.

 Chapter 8 is the last chapter and it presents the conclusions obtained from this master thesis.

(11)

2. Air Cooled Blast Furnace Slag (Hyttsten)

2.1. General

Blast Furnace Slag (Masugnsslagg) is an alternative construction material that was used for a long time in Sweden as well as in other countries for roads, railways, grounds, etc. It is used for Base, Sub- base layers and for other filling works in road and building construction works (Lindberg, 2004). Blast Furnace is one of the cheapest and easily accessible materials (Persson, 2004). Using Blast Furnace Slag saves the natural resources of gravel and crushed stones. The interest in using such alternative aggregate materials in Sweden is to achieve the Road Administrations’ objectives in saving the natural resources also to achieve the government’s environmental objectives “Build on a good Environment”

(Andersson, 2005).

Blast Furnace Slag (Masugnsslagg) is a byproduct of iron production. Heating iron ore, coke (as fuel) and limestone (as slag former) in the Blast Furnace produces raw iron (pig iron) and Blast Furnace Slag, which is divided into two different groups according to the cooling methods:

Air Cooled Blast Furnace Slag (Hyttsten); is cooled slowly by air and has a crystalline texture, which is crushed and sieved into different fractions of Air Cooled Blast Furnace Slag.

Water Cooled Blast Furnace Slag or Granulated Slag (Hyttsand); is cooled rapidly in water and has a non- crystalline texture (Andersson, 2005).

The amount of the produced Slag depends on the iron content in the ore, the lower iron content the higher Slag is produced (Lindberg, 2004).

Air Cooled Blast Furnace Slag (Hyttsten) is used as a road construction material since 1970’s for Sub- base layers, Base layers of streets, pedestrian and bicycle- roads. It is even used as a slip resistance material in roads in Norrbotten, while the Granulated Blast Furnace Slag (Hyttsand) having non- crystalline texture is not used in road construction; however it is used as stabilizer of Base layer and as a filling material (Persson, 2004).

Today the company that produces and sales Air Cooled and Granulated Slag in Sweden is SSAB. The companies Blast Furnaces in Luleå and Oxelösund produce annually about 500,000 tons Blast Furnace Slag (Masugnsslagg), 6000,000 tons of Air Cooled Blast Furnace Slag (hyttsten) have been used around Luleå for the period 1976- 2006 (Persson, 2004; Ullberg, Rutqvist, & Sandström, 2005).

A part of Air Cooled Blast Furnace Slag is available as Unsorted Air cooled Blast Furnace Slag (was called fraction 0- 300mm), and the rest are crushed and sieved into two fractions 0-30mm and 0- 100mm. One ton Air cooled Blast Furnace Slag costs around 40- 50kr (transportation included) depending on the distance and the quantity, and since its density is 1.3 t/m³, each m³costs about 65kr.

While one ton natural gravel costs 60- 70kr, and due to its density, which is 1.8 t/m³, each m³costs about 108kr (Kanschat, 1996).

2.2. Properties of Air Cooled Blast Furnace Slag

Air Cooled Blast Furnace Slag’s chemical composition differs according to its producing process.

While the geometrical and physical properties are somehow the same, which gives the same technical properties (Andersson, 2005).

2.2.1.Chemical Properties

2.2.1.1. Chemical Composition

Air Cooled Blast Furnace Slag’s main components are Calcium oxide (CaO), Silicon dioxide (SiO₂), Magnesium oxide (MgO) and Aluminum oxide (Al₂O₃). Its composition depends on the raw materials (iron ore, coal and limestone) and the procedure in the Blast Furnace (Andersson, 2005).

(12)

CaO/SiO₂ratio is used to know either the Slag is acidic or basic, if CaO/SiO₂< 1 then the Slag will be acidic otherwise it will be basic. The Air Cooled Blast Furnace Slag from SSAB in Luleå is acidic i.e.

SiO₂content is higher than CaO (Kanschat, 1996).

Table 2.1 shows the mean values of the main continents in Air Cooled Blast Furnace Slag, Gabbros¹ and Moraine² collected from Kanschat, 1996 that compared Air Cooled Blast Furnace Slag with the other two natural materials to show the difference in their continents.

Table 2.1 The main elements in Air Cooled Blast Furnace Slag, Gabbros and Moraine

The data is collected from Kanschat, 1996 compared with Persson, 2004 and the analysis made for SSAB Tunnplåt AB at 2007-09-03 by LU_App1 (see Appendix 1).

Elements Unit Air Cooled Blast

Furnace Slag Gabbros Moraine

SiO2 % 33.700 34.000 54.700 64.300

Al2O3 % 9.660 13.200 17.000 14.100

CaO % 32.500 32.800 6.060 3.660

Fe2O3 % 0.320 1.390 12.000 6.630

K2O % 0.710 0.800 1.050 2.720

MgO % 15.200 18.100 5.000 2.120

MnO % 0.365 0.460 0.180 0.090

Na2O % 0.558 0.670 3.030 3.900

P2O5 % 0.008 0.020 0.300 0.210

TiO2 % 2.130 2.190 1.410 0.980

As (Arsenic) ppm <0.100 0.730 11.200 0.930

Ba (Barium) ppm 376.000 767.000 868.000 685.000

Be (Beryllium) ppm 5.920 8.050 2.780

Cd (Cadmium) ppm <0.010 0.090 0.117

Co (Cobalt) ppm <1.000 2.020 24.300 13.000

Cr (Chrome) ppm 29.100 84.500 238.000 75.000

Cu (Copper) ppm 0.966 <1.000 70.500 35.000

Hg (Mercury) ppm <0.05 0.070 0.146 0.045

La (Lanthanum) ppm 69.400 76.100 36.200 35.000

Mo (Molybdenum) ppm <5.770 <6.000 <5.030

Nb (Niobium) ppm <5.770 9.840 11.800 14.000

Ni (Nickel) ppm <1.000 2.530 110.000 23.000

Pb (Lead) ppm <0.100 0.280 2.750 3.900

Sc (Scandium) ppm 28.000 34.900 24,500 15.000

Sn (Tin) ppm <20.000 71.500 <20.100

Sr (Strontium) ppm 290.000 481.000 480.000 265.000

V (Vanadium) ppm 416.000 572.000 215.000 106.000

W ppm <23.300 <60.000 <20.100

Y (Yttrium) ppm 44.700 60.100 15.700 30.000

Yb (Ytterbium) ppm 5.680 3.090 3.700

Zn (Zink) ppm 1.1300 2.860 149.000 45.000

Zr (Zirconium) ppm 201.000 293.000 173.000 408.000 S (Sulfur) ppm 9720.000 13300.000 3000.000

The following strange values were found in Persson, 2004:

(Cu=235ppm, La=0.07ppm, Ni=8.41ppm, Pb=4.80ppm, W<572ppm and Zn=115ppm). And the percentage of SiO2was less than CaO in Persson, 2004.

The trace metals³, that exist in Air Cooled Blast Furnace Slag in higher concentrations compared with Gabbros and Moraine in Norrbotten are (Be, La, Sc, Sn, V, Y and Yb) (Kanschat, 1996). Vanadium (V) content (416- 572 ppm) in Air Cooled Blast Furnace Slag is higher than the Landscape Protection

1Gabbros are from Svalgets Bergtäkt in Boden.

2Moraine is from Kalix river valley at Norrbotten.

3Trace metals are the metals found in animals and plants cells and tissues, which include (Fe, Mg, Z, Cu, Ni, Co, V, As, and Mo). They normally exist in low concentrations.

(13)

Agency’s standards for KM⁴, MKM⁵and MKM GV⁶(Persson, 2004). And it is two times higher than Gabbros and four times higher than Moraine in Norrbotten (Kanschat, 1996).

The heavy metals⁷that are more considered in Environment issues are -As, Cd, Co, Cr, Cu, Hg, Ni, Pb and Zn-. The concentrations of these metals are low in Air Cooled Blast Furnace Slag compared with Gabbros and Moraine (Kanschat, 1996).

The Sulfur content in the Air Cooled Blast Furnace Slag is very high compared to Gabbros and Moraine.

2.2.1.2. Chemical Weathering

Air Cooled Blast Furnace may be weathered by Lime and Iron weathering. Lime stone weathering occurs especially during the first days when the Slag is a strong basic. The Slag from Luleå seems to be acidic and its CaO/SiO₂ration is less than 1, thus the Lime weathering may not happen (Lindberg, 2004).

Iron weathering the partly reduced Iron oxide which’s volume, will increase by further oxidation. The Air Cooled Blast Furnace Slag, which can be weathered easily, is known within two days of water stratification. The probability of Iron weathering of today’s well controlled Slag is very low (Lindberg, 2004).

2.2.1.3. Leachable elements

Leaching means that the aggregate material badly weathers or reacts with water and certain elements will dissolve. These dissolved elements may be transported through percolation or by water stream down to the ground water or to the surface water (Persson, 2004).

The amount of the trace metals leaching from old Air Cooled Blast Furnace Slag and the other natural materials are equal. It leaches less heavy metal than Gabbros and Moraine (Kanschat, 1996). Sulfur is the most leachable elements for Air Cooled Blast Furnace Slag, while in the natural materials Sulfur remains in the material (Tossavainen & Forssberg, 1999).

Table 2.2 shows the availability for leaching in Air Cooled Blast Furnace Slag and one of the natural materials used by Tossavainen & Forssberg, 1999.

Table 2.2 The availability for leaching of Sulfur and other elements in Air Cooled Slag and Gabbros. From (Tossavainen & Forssberg, 1999).

elements

Air Cooled Blast Furnace Slag Gabbros Total quantity

(ppm) Availability for

leaching % Total quantity

(ppm) Availability for leaching %

S 9720.00- 13300.00 21.40 2000.00 2.30

V 416.00- 572.00 6.48 282.00

Cr 29.10- 84.50 0.86 161.00 0.19

Cu <1.00 56.30 0.40

Ni <1.00- 2.53 29.20 6.90

Pb <0.10- 0.28 6.50 6.30 0.40

Zn 1.13- 2.86 48.80 75.70 3.40

The total quantities for Air Cooled Blast Furnace Slag are from table 1.1.

4Känslig markandanvändning- Sensitive land uses 100mpp.

5Mindre känslig marknadanvändning- Less sensitive land uses 200mpp.

6Mindre känslig marknadanvändning med grundvattenskydd- Less sensitive land uses with ground water protection 200mpp.

7Heavy metals definition is that its density exceeds 5g/cm³. They include metallic elements 21 to 84 from Sc to Po. The heavy metals that are important from environmental view point are (As, Pb, Cd, Co, Cu, Cr, Hg, Ni, Sn, V and Zn)

(14)

2.2.1.3.1. Sulfur (S)

Air Cooled Blast Furnace Slag contains 1-2% Sulfur. The high Sulfur content may cause leaching Sulfur in Sulfur compounds such as Sulfates. At the same time a certain amount of acids are produced from oxidizing of Sulfur to Sulfates due to liberation of Hydrogen ions. Air cooled Blast Furnace Slag may smell sulfur (Andersson, 2005).

Oxidizing of reduced Sulfur compounds is illustrated by that Hydrogen Sulfide reacts with acids and water and is oxidized into Sulfates; the reaction produces also Hydrogen ions (Kanschat, 1996).

2.2.2.Physical Properties

A material’s applicability for road construction is decided to a great extent according to its physical properties. Some of the most important physical properties of Air Cooled Blast Furnace Slag and Rock materials are shown in table 2.3 (Ullberg, Rutqvist, & Sandström, 2005).

Table 2.3 Physical properties of Air Cooled Blast Furnace Slag and Rock materials from (Ullberg, Rutqvist, & Sandström, 2005).

Materials Property Air Cooled Blast

Furnace Slag Rock Material Thermal Conductivity (W/M.K) 0.30- 0.60 1.30- 1.60

Porosity (%) 39- 46 25

Solid Density (t/m³) 2.90- 3.10 2.60- 3.10

Bulk Density (t/m³) 2.2*

0-30mm 1.35

0-90mm 1.32

Unsorted 0-300mm 1.20

E- Module (MPa) 300- 600 300- 600

Micro Deval 15- 39 <20

Hydraulic Properties Yes No

* In Kanschat, 1996 the value is1.7 t/m³,then it was corrected by the supervisor of this thesis (prof. Sven Knutsson)

2.2.2.1. Thermal Conductivity

Air Cooled Blast Furnace Slag has low thermal conductivity. A low thermal conductivity value means that it has a good thermal insulation, a property that can be considered as an advantage in road construction (Lindberg, 2004). Experiences show that 90cm of Air Cooled Blast Furnace Slag of fraction 0-100mm gives the same thermal insulation as 8cm of Styrofoam. Using Air Cooled Blast Furnace Slag in road construction will prevent the frost to go deep (Kanschat, 1996).

A disadvantage of having good thermal insulation is that it will prevent heat from going down through the road body in spring (Kanschat, 1996). If a material with low thermal conductivity is used as an upper layer in a roads body it may cause risks of frost slipperiness on the surface. Thus there are restrictions in Sweden on using Air cooled Slag and other materials in road construction (Lindberg, 2004).

2.2.2.2. Porosity

Porosity is the pore volumes divided by the total volume. Air Cooled Blast Furnace Slag has high porosity. This high porosity is mainly due to the Sulfur content in the melted slag as Sulfides are oxidized to Sulfur dioxide and emitted in gas form, which will form gas blisters by cooling. The gas blisters may be big but they are not connecting with each other. This makes the material despite having high porosity, not to suck a lot of water, which means it will have low capillarity (Lindberg, 2004).

(15)

2.2.2.3. Density

Air cooled Slag Blast Furnace Slag’s compact density is almost the same as the rock materials. Due to the high porosity its bulk density will be considerably low (Lindberg, 2004). Air Cooled Blast Furnace Slag’s bulk density is 1.2-1.4 t/m³, compared to the bulk density of crushed rock which is about 1.7 t/m³. This means that the weight of the same volume of Air Cooled Blast Furnace Slag will be 20- 30% less than crushed stone’s (Kanschat, 1996).

Road embankments of light masses can minimize or eliminate the demand of pressure embankment and strengthening arrangements that will decrease the width of the road and reduces the costs. To use a material with low bulk density is favorable as foundation disposition may occur on settling sensitive underground and /or where stability problems exist (Lindberg, 2004).

2.2.3.Mechanical properties

The most important mechanical properties that are considered in using unbound aggregate materials in road construction are:

- Stiffness number, which is a measure of the resilient deformation’s resistance of a material, it can be expressed as E- module⁸.

- Stability, which is a measure of the permanent deformation’s resistance of a material.

- Bearing capacity is the load that a layer can carry without being deformed more than allowed.

These properties depend on the particle size, particle shape, its distribution and the mineral composition (Arm, 2003).

Air Cooled Blast Furnace Slag can be used without problems in base level as well as for sub-base layer. The research showed that Air cooled Blast Furnace Slag has higher stiffness than crushed stone and natural materials at the stress level that is actually in the sub-base layer. Many reports concluded that the stiffness and stability properties make the Air cooled Blast Furnace Slag replace the conventional materials used for base and sub-base layers. Experiments showed that the stiffness of Air Cooled Blast furnace Slag will be increased by time as it will be crushed. The fine materials in Slag behave like a cemented substance by excess of water, while it behaves like a lubricating staff in natural materials (Lindberg, 2004).

8Young’s Modulus also is known as modules of elasticity, used to measure the stiffness.

(16)

3. Percostations

3.1. General

Percostations are measuring techniques developed by Estonia Adek Ltd, to be used in detecting critical bearing capacity of road materials and to assess the frost and water effects on the road materials. It measures and plots dielectric constant (Dielectricity), Electrical Conductivity and Temperature, figure 3.1showes a sample of percostation data collected from the test road in Luleå (Luleå percostation).

The figure contains 3 graphs which represent Dielecriciry values, Electrical conductivity and Temperature respectively. The different colors indicate the percomter sensors that are installed at different depths. The small diagram to the right represents the percostation, and it shows the depth and the color of each percomter.

Figure 3.1 a sample of the prcostation data collected from Luleå percostation at 22.05.2008. Each color represents different sensors at different depths.

According to Saarenketo, 2006 a the percostation consists of:

- Percometers, which are sensors installed in the road section at different depths. A percometer can measure dielectric values, electrical conductivity and temperature from eight channels.

Usually the measurements are collected each 2 hours.

- Data collector, which will be located on the road side, and used to store the data.

- Data transfer and analyzing system, which will transfer the readings through a GSM or telephone network to the base where the data can be stored in files. Each file contains information on the dielectric constant (dielectricity), electrical conductivity, and temperature also it gives information on the structure, used materials, calculated critical tension for each station.

- Solar panels, to provide power.

A Percostation system diagram is illustrated in figure 3.2 which consists of five sensors and the data collector. The system generates its own electricity power through a solar panel and transfers data by a General Packet Radio Services- GPRS- system.

(17)

Figure 3.2 a percostation system from (Saarenketo 2006 b); the weather station is removed from the diagram.

3.2. Electrical properties of road materials

The relation between electrical properties and the permanent deformation of road materials were studied for the last 10 years and it was found that there is a strong relation between them (Saarenketo, 2006 a).

The main components of road materials are; air, water/ ice, minerals and the weathering products, clay, colloid particles, organic material and salts. As mentioned in item 3.1, the most important electrical properties that are monitored by the Percostations are Electrical Conductivity and the Dielectric Permittivity (Saarenketo, 2006 a).

3.2.1.Electrical Conductivity

The electrical conductivity of a material is the measure of its ability to conduct an electric current, it is affected by the water content, meniral quality, ion content, colloids and the temprature. Electrical conductivity’s unit is µS/cm or µmho/cm and 1 µS/cm=1 µmho/cm⁹. Most of non conducters such as quartz, micas and feldspars can be changed to electrical conducters when it comes in contact with water. Electrical conductivity monitored by the Percostations indicate the salt content or total ionic concentrations in the water. It provides information concerning pore water (Saarenketo, 2006;

Hänninen, 2001).

3.2.2.Dielectricity (Dielectric Constant or Dielectric permittivity)

Dielectric permittivity is the most important electrical property that is registered by the Percostations.

Since the road materials and sub-grades are composite materials, their dielectric values will be a combination of the individual dielectric constants of each of these components, their volume fraction and geometrics (Saarenketo, 2006 a). The Dielectric permittivity of soil and road materials is affected by the water content, which is affected by the mechanical properties of the material. The more fine particles, the higher water content (Hänninen, 2001). Dielectric values are used to assess the quality of unbound and bound aggregates used in road construction. The dielectric permittivity of materials in the nature varies between 1 for the air to 81 for polar water at 20^oC (Saarenketo, 2006 a).

9Siemens (s) was used since 1970’s as an International System of Units (SI) derived unit of electrical

conductivity to replace the previously used unit (mho). The term (mho) was used as an inverse fro ohm (Ω), thus each Siemens equals to 1/ Ω.

(18)

4. The test road

The test road is a part of the road 563 between Sjulsmark and Ersnäs. When about 16 Km was rebuilt in August- September 2006, 800 m was built as a long term test road by using Air Cooled Blast Furnace Slag. The project was done in cooperation between the Road Administration Northern region –Vägverket-, SSAB Tunnplåt, SSAB Merox and BDX to:

- Assess using Air Cooled Blast Furnace Slag - Hyttsten- in road construction and compare it with rock materials during the construction and in long term.

- Modify and update the “Common Technical Description for Road Building”- “ATB Hyttsten”¹⁰-.

The road was divided into four parts of 200m length, three were constructed by using three different fractions of Air Cooled Blast Furnace Slag, which were unsorted, 0/125, 0/63 and the fourth part was constructed by using normal crushed rock. Each part was treated by two different kinds of rollers, 100m by using ordinary roller and the other 100m by ring roller (see figure 4.1.a). The same process was applied in construction of all the subsections (Ullberg, Rutqvist, & Sandström, 2005).

A percostation system was installed for subsection 5, which was constructed by Air Cooled Blast Furnace Slag using ring roller. The percometers - sensors- were installed at 15cm, 30cm, 50cm, 80cm and 110cm depth.

Figure 4.1.a shows a diagram of the test road

Figure 4.1.b shows a picture of the test road and the percostation location

10ATB Väg and ATB Hyttsten are the Swedish Road Administration’s technical documents

(19)

5. Data analysis

5.1. Data sources

To analyze interrelationships between the variables (Dielctricity, Electrical Conductivity and temperature) and to investigate the differences between those variables at Luleå Percostation where Air cooled Slag was used and the other two Percostations. Data was obtained from the following resources:

1. Roadscanners, data was provided by Roadscanners for the Percostations at:

- Inari Percostation in Finland for the period 2006-09-06 to 2008-04-10.

- Luleå Percostation in Sweden for the period 2006-08-23 to 2008-04-10 then to 2008-05- 10.

- Koskenkylä Percostation in Finland for the period 2005-11-22 to 2006-09-07.

Figure 5.1 a map shows data collection positions at Luleå, Inari and Koskenkylä Percostatios.

The obtained data was on:

- The Electrical Conductivity values.

- The Dielectricity values.

- The Temperature.

2. Swedish Meteorological and Hydrological Institute- Environment & Safety Services on:

- The temperature and the Humidity at Luleå Flygplats (Air port) station for the period 2006-09-01to 2008-05-12.

- The precipitation at Luleå Bergnäset station for the period 2006-09-01 to 2008-05-11.

5.2. Descriptive Statistics

The data provided by the Percostations contain 16 variable divided into three sets, which are Dielectricity values for the ports (P0, P1, P2, P3 and P4), Electrical Conductivity values for the ports (P0, P1, P2, P3 and P4) and the Temperature for the ports (P0, P1, P2, P3, P4 and P5) in addition to

(20)

one categorical variable, which is the three stations (Luleå, Inari and Koskenkylä). There are 15933 cases divided between the stations as following:

- Luleå 5782 cases, representing 36.3%.

- Inari 6746 cases, representing 42.3%.

- Koskenkylä 3405 cases, representing 21.4%.

The data provided by Swedish Meteorological and Hydrological Institute- Environment & Safety that were collected from stations around Luleå are divided into two different sets. One set contains two variables, which are (Temperature and Humidity) and it consists of 4947 cases. The second set contains only one variable, which is Precipitation and it has 1231 cases.

5.2.1.Checking the errors and missing data

Before doing any statistical analysis it is important to check the data for the input errors, missing values and to find out any violation to these assumptions (Pallant, 2005). It was found that:

1. No input errors were found.

2. The following data recording gaps are investigated (see Appendix 2).

- Luleå Percostation; (09 Sep 2006- 11 Sep 2006, 05 Oct 2006- 18 Oct 2006, 07 Jan 2007- 19 Mar 2007, 12 Sep 2007- 30 Sep 2007, 01 Oct and 07 Oct 2007- 12 Oct 2007, 02 Jan 2008- 06 Jan 2008).

- Inari Percostation; (17 Dec 2007- 23 Dec 2007, 23 Jan 2008- 31 Jan 2008).

- Koskenkylä Percostation; (01 Feb 2006- 02 Feb 2006).

3. All the variables in Luleå Percostation have missing cases, but they are very high for the variable Dielectricity value- Port 2 (15%) and Port 3 (48.7%). Regarding the data provided by Swedish Meteorological and Hydrological Institute- Environment & Safety, there are only 4 missing cases for humidity, 3 cases in precipitation and no missing cases in temperature. Inari Percostation has no missing values. While Koskenkylä Percostation has missing values at Dielectricity value- Port 0 (86.9%), Dielectricity value- Port 3 (32.2%) and Temperature- Port 4 (100%). The details are presented in table 4.2 (see Appendix 3).

5.2.2.Checking the outliers

Outliers are the out of range cases. Most of the statistical analysis techniques are sensitive for outliers or extreme values, which may affect means, variances, covariance, correlation etc. Different tools are available to identify and check the outliers (Pallant, 2005). “Box-plot" is used in this report to identify the outliers, and to assess the effect of outlier cases obtained mean values are compared with 5%

Trimmed Mean.

“Box-plot” is a tool to identify outliers. The cases are identified as outliers (indicated with circles) if they extend more than 1.5 box-lengths from the edge of the box, and the extreme cases (indicated with asterisks) are the cases, which extend more than 3 box-lengths from the edge of the box (Pallant, 2005). The box-plots of all the variables are shown in (Appendix 4). The Box-plots show that most of outliers and extreme values were in Luleå Percostation in the variables Dielectricity values at ports (1, 2, 3 and 4), Electrical Conductivity at ports (0, 1, 2, 3 and 4). At Inari Percostation the box-plot show outliers in the variables Dielectricity value at ports (1 and 4) and Electrical Conductivity at ports (0 and 1), while Koskenkylä Percostation has outliers in Dielectricity values port 3.

To check outliers and extreme values affect, 5% Trimmed mean values (obtained by removing the top and bottom 5% of the cases and recalculating a new mean value) are compared with the original mean values and it is found that the outliers and extreme values for most of the variables that was found to be outliers in box-plot have not big influences on the mean see table 5.2. Only the variables Electrical Conductivity (Port 1 and Port) in Luleå Percostation were most affected by the outliers. The variables Electrical Conductivity P 1 and P2 were more inspected and it was found that:

- The highest difference between the original mean and 5% Trimmed mean was in 2008.

- The difference was due to the seasonal changes and during thaw periods.

(21)

- Include a big number of the cases. And removing or changing all the extreme values will affect the other cases and variables.

Thus all the data was retained.

The same process is followed with the precipitation, humidity and temperature data that was provided by Swedish Meteorological and Hydrological Institute- Environment & Safety. After checking box- plots (see Appendix 4) and the influence of the outliers by checking 5% Trimmed mean values (see table 5.1) all the data were retained.

Table 5.1 checking the data provided by the Swedish Meteorological and Hydrological Institute- Environment & Safety. The minimum, maximum, mean and 5% Trimmed mean values of each variable are explored.

Luleå

Precipitation Humidity Temperature

Total N. 1231 4947 4947

Number 1228 4943 4947

Missing 3 4 0

Min. value 0,00 19,00 -28,70

Max. value 27,80 99,00 26,200

Mean 0,97 82,35 1,98

5% T. mean 0,51 83,70 2,13

5.2.3.Assessing Normality

Most of the statistical techniques assume that the distribution of data must be Normal; to assess the normality of the distribution of data the following tools are used (see Appendix 5):

- ‘Test of Normality’ table available by SPSS, which gives Kolmogrove- Smirnov statistic to assess the normality (see appendix 5.1). According to Pallant, 2005, a non significant result (Sig. values of more than 0.05) indicates normality. The sig. values were less than 0.05 for most of the cases, thus the distribution cannot be considered as normal distribution in this stage.

- Then ‘Histograms’ for each case is checked, again they show violation to the normality assumption.

- Finally the normal probability plot ‘Normal Q-Q Plots’ is inspected, which plots the observed value of each case against the expected value from the normal distribution. Most of the cases were found to have non-normal distribution.

Since non significant results were obtained from transforming of the data. It was decided to retain the data in its original form.

The assumptions linearity that assumes the relation between the variables should be a straight line and homoscedsticity that assumes the scores for each variable should be similar at all values of any other variable, were checked by inspecting the scatter-plots, which is available in SPSS (see Appendix 5).

Many variables show violation to the assumptions normality and linearity, which should be considered when the correlation analysis is performed.

5.3. Correlation analyses

Correlation analyses were performed by using Pearson product-moment, which is provided by SPSS to investigate interrelationships between the variables per each Percostation (see Appendix 6). The Correlation coefficients or Person correlation values indicated in the tables show the strength of relationship between the variables and the negative sign refers to the direction of these relationships.

Significance levels are checked too, and having a big sample all the correlations reach the statistical significance at (p<0.05).

Generally it is noticed in the tables:

- Different values of correlation coefficient indicate different levels of interrelationships.

- Different signs (positive and negative) indicate different directions of relationships.

(22)

Table 5.2 checking the data provided by the Percostations. First total data set is checked for all the 16 variables and the categorical variable (Stations). Then the same procedure is provided for each Percostations (Luleå, Inari and Koskekylä). The minimum, maximum, mean and 5% Trimmed mean values of each variable are also explored in the table.

Stations Dielectricity (Er) Value Electrical Conductivity Temperature

P 0 P 1 P 2 P 3 P 4 P 0 P 1 P 2 P 3 P 4 P 0 P 1 P 2 P 3 P 4 P 5

Total

Total N. ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ ¹⁵⁹³³ Number ¹⁵⁹³³ ¹²⁹⁵⁷ ¹⁵⁹¹² ¹⁵⁰⁶⁸ ¹²⁰¹⁹ ¹⁵⁹¹⁰ ¹⁵⁹¹³ ¹⁵⁹¹³ ¹⁵⁹⁰⁵ ¹⁵⁹¹¹ ¹⁵⁸⁸⁸ ¹⁵⁸⁸¹ ¹⁵⁸⁸⁵ ¹⁵⁸⁸² ¹⁵⁸⁸⁴ ¹²⁴⁸¹ ¹⁵⁹¹⁸

Missing ⁰ ²⁹⁷⁶ ²¹ ⁸⁷⁵ ³⁹¹⁴ ²³ ²⁰ ²⁰ ²⁸ ²² ⁴⁵ ⁵² ⁴⁸ ⁵¹ ⁴⁹ ³⁴⁵² ¹⁵

Min. value ¹ ^2,51 ^3,27 ^1,03 ^1,02 ^2,42 ^,00 ^,00 ^,00 ^,00 ^,00 ^-23,40 ^-19,00 ^-17,00 ^-13,40 ^-1,70 ^-32,2

Max. value ³ ^17,30 ^34,60 ^36,70 ^25,80 ^23,50 ^129,00 ^431,00 ^354,00 ^437,00 ^465,00 ^37,10 ^23,60 ^22,00 ^19,00 ^16,70 ^48,70

Mean ² ^5,75 ^6,84 ^11,48 ^12,35 ^10,64 ^4,21 ^13,42 ^13,74 ^58,28 ^42,71 ^2,54 ^2,82 ^3,13 ^3,25 ^3,56 ^3,35

5% T. mean

Luleå

Total N. ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸² ⁵⁷⁸²

Number ⁵⁷⁶⁴ ⁵⁷⁶¹ ⁴⁹¹⁷ ²⁹⁶⁴ ⁵⁷⁶⁶ ⁵⁷⁶² ⁵⁷⁶² ⁵⁷⁵⁴ ⁵⁷⁶² ⁵⁷³⁷ ⁵⁷³⁰ ⁵⁷³⁴ ⁵⁷³¹ ⁵⁷³³ ⁵⁷³⁵ ⁵⁷⁶⁷

Missing ¹⁸ ²¹ ⁸⁷⁵ ²⁸¹⁸ ¹⁶ ²⁰ ²⁰ ²⁸ ²⁰ ⁴⁵ ⁵² ⁴⁸ ⁵¹ ⁴⁹ ⁴⁷ ¹⁵

Min. value ^4,81 ^3,71 ^1,03 ^1,02 ^5,29 ^,00 ^,00 ^,00 ^1,00 ^2,00 ^-18,90 ^-8,90 ^-7,60 ^-3,10 ^-1,70 ^-19,60

Max. value ^17,30 ^34,60 ^36,70 ^19,70 ^23,50 ^129,00 ^431,00 ^354,00 ^437,00 ^465,00 ^37,10 ^23,30 ^22,00 ^16,70 ^16,70 ^41,50

Mean ^7,28 ^7,21 ^4,95 ^9,69 ^10,83 ^8,63 ^23,95 ^10,05 ^105,10 ^113,53 ^4,90 ^4,20 ^3,99 ^4,03 ^4,43 ^5,12

5% T. mean ^6,62 ^4,28 ^9,62 ^10,58 ^7,54 ^15,71 ^4,84 ^97,06 ^105,75 ^4,71

Inari

Total N. ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶

Number ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶ ⁶⁷⁴⁶

Missing ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰ ⁰

Min. value ^2,51 ^3,27 ^6,41 ^5,58 ^5,25 ^,00 ^,00 ^,00 ^,00 ^,00 ^-19,30 ^-12,80 ^-5,20 ^-3,50 ^-1,40 ^-28,10

Max. value ^9,41 ^15,70 ^30,40 ^25,80 ^18,80 ^4,00 ^5,00 ^29,00 ^37,00 ^6,00 ^25,20 ^20,90 ^15,10 ^13,50 ^11,40 ^44,60

Mean ^4,50 ^6,05 ^17,63 ^16,50 ^13,72 ^,02 ^,22 ^12,78 ^17,51 ^3,31 ^,65 ^1,41 ^2,37 ^2,55 ^2,83 ^1,08

5% T. mean ^5,80 ^13,94 ^,00 ^,13 ^,70

Koskenkylä Total N. ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵

Number ⁴⁴⁷ ³⁴⁰⁵ ³⁴⁰⁵ ²³⁰⁹ ³³⁹⁸ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰³ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ³⁴⁰⁵ ⁰ ³⁴⁰⁵

Missing ²⁹⁵⁸ ⁰ ⁰ ¹⁰⁹⁶ ⁷ ⁰ ⁰ ⁰ ² ⁰ ⁰ ⁰ ⁰⁰ ⁰ ³⁴⁰⁵ ⁰

Min. value ^4,76 ^4,35 ^3,85 ^1,55 ^2,42 ^,00 ^,00 ^,00 ^,00 ^,00 ^-23,40 ^-19,00 ^-17,00 ^-13,40 ^-32,20

Max. value ^5,21 ^14,50 ^17,00 ^4,98 ^7,15 ^12,00 ^53,00 ^57,00 ^147,00 ^4,00 ^25,60 ^23,60 ^21,60 ^19,00 ^48,70

Mean ^4,94 ^7,80 ^8,71 ^3,62 ^4,22 ^5,03 ^21,76 ^21,88 ^59,83 ^1,45 ^2,30 ^3,28 ^3,17 ^3,31 ^4,86

5% T. mean ^3,63 ^4,76

(23)

- Since there is a big sample, most of the correlations reach the statistical significance level according to Pallant, 2005.

And since the data provided by Swedish Meteorological and Hydrological Institute- Environment &

Safety are in separate files and have its special dating and timing format, thus daily mean values of Dielectricity, Electrical Conductivity and Temperature was measured and listed against total daily values of precipitation then correlation analyses were performed to the new set of data for Luleå Percostation (see Appendix 6 – table 2).

5.4. Numerical analyses

The data was numerically anlysed by calculating maximmum, minimum and mean values for all the variables (from Percostations and from the Swedish Meteorological and Hydrological Institute- Environment & Safety) and it was summarized in tables (see Appendix 7).

5.5. Visual analysis

Numerical values obtained from the data sources are visialy explored by using different types of graphes, which are provided by Microsoft Excel. The data was visualy explored in different modes to investigate the sesonal changes of the variables in each Percostation, the differences and the similaritys between the Percostation’s variables and the changes through the Ports, which represent the different layers or materials in the road body (see Appendix 8 and Appendix 9). In order to inspect the seasonal changes, the data was explored in monthly bases for each Percostation;

- Luleå Percostation for the period August- 2006 to May- 2008.

- Inari Percostation for the period September- 2006 to April- 2008.

- Koskenkylä Percostation for the period November- 2005 to September 2006.

6. Results

6.1. Correlation analyses

Following results were obtained from the correlation analyses:

1. Luleå Percostation;

- Only the Dielectricity value in Port 0- the base layer- has correlation with the Temperature.

As shown in appendix 6- table 1. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity in Port 0- the base layer- and the Temperature generally.

- The Dielectricity values in Port 3 and Port 4- the sub grade layer- has a weak correlation with Temperature. The numbers shown in appendix 6- table 1 are under 0.3, which represent an inconsiderable correlation.

- Only Electrical Conductivity in Port 3 and Port 4- the sub grade layer- has correlation with the Temperature. As shown in appendix 6- table 1. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Electrical conductivity in Port 3 and Port 4- the sub grade layer- and the Temperature generally.

- The correlation between Dielectricity value and Electrical Conductivity is strong always at the same port. As shown in appendix 6- table 1. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity and the Electrical conductivity in the same port.

- There is a strong correlation (over 0.70) between Dielectrity values in Port 1 and Port 2 - the sub base layer- and between Port 3 and Port 4- the sub grade layer-. As shown in appendix 6- table 1. The numbers higher than 0.70, which are in gray, indicate a very high correlation

(24)

between the Dielectricity in Port 1 and Port 2 - the sub base layer-, and the Dielctricity values in Port 3 and Port 4- the sub grade layer-.

- There is a strong correlation (over 0.70) between the Electrical Conductivity in Port 1 and Port 2 - the sub base layer- and between Port 3 and Port 4- the sub grade layer-. As shown in appendix 6- table 1. The numbers higher than 0.70, which are in gray, indicate a very high correlation between the Electrical conductivity in Port 1 and Port 2 - the sub base layer-, and the Electrical conductivity values in Port 3 and Port 4- the sub grade layer-.

The analyses show that there is no correlation between Precipitation and the other variables, unless a very small correlation between Precipitation and the Dielectricity values in the sub grade layer Port 3 (0.115) and Port 4 (0.163), and between Precipitation and Electrical Conductivity in the sub grade layer Port 3 (0.137) and Port 4 (0.134) as shown in Appendix 6- table 2 in blue.

2. Inari Percostation;

- The Dielectricity values in all the layers (Port 0, Port 1, Port 2, Port 3 and Port 4) have correlation with the Temperature. As shown in appendix 6- table 3. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity in all the layers (Port 0, Port 1, Port 2, Port 3 and Port 4) and the Temperature generally.

- The correlation between Dielectricity value and Electrical Conductivity is strong in the same port, except Port 0- the base layer-. As shown in appendix 6- table 3. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity and the Electrical conductivity in the same port, except Port 0- the base layer-. Where no correlation between Dielectricity and Electrical conductivity is shown.

- There is a strong correlation (over 0.70) between Dielectrity values in P0 and P1 and between P2, P3 and P4. As shown in appendix 6- table 3. The numbers higher than 0.70, which are in gray, indicate a very high correlation between the Dielectricity in Port 0- the base layer- and Port 1- the sub base layer- , and the Dielctricity values in Port 2- the protection layer- , Port 3 and Port 4- the sub grade layer-.

- There is a strong correlation (over 0.70) between the Electrical Conductivity in Port 0- the base layer- and Port 1- the sub base layer- and between Port 2- the protection layer- , Port 3 and Port 4- the sub grade layer-. As shown in appendix 6- table 3. The numbers higher than 0.70, which are in gray, indicate a very high correlation between the Electrical conductivity in Port 0- the base layer- and Port 1- the sub base layer-, and the Electrical conductivity values in Port 2- the protection layer- , Port 3 and Port 4- the sub grade layer-.

3. Koskenkylä Percostation;

- The Dielectricity values in all the layers (Port 0, Port 1, Port 2, Port 3 and Port 4) have correlation with the Temperature; the correlation is negative in Port 0- the base layer- and Port 3- the sub grade layer-. As shown in appendix 6- table 4. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity in all the layers (Port 0, Port 1, Port 2, Port 3 and Port 4) and the Temperature generally. The negative sign in Port 0- the base layer- and Port 3- the sub grade layer-, indicates the correlations direction.

- The Electrical Conductivity in all the ports layers (Port 0, Port 1, Port 2, Port 3 and Port 4) has correlation with the Temperature. As shown in appendix 6- table 4. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Electrical conductivity in all the layers and the Temperature generally.

- The correlation between Dielectricity value and Electrical Conductivity is strong in the same port, except in Port 0- the base layer- and Port 3- the sub grade layer-. As shown in appendix 6- table 4. Numbers between 0.30- 0.70, which are in yellow represent a reasonable correlation between the Dielectricity and the Electrical conductivity in the same port, except the Port 0- the base layer- and Port 3- the sub grade layer-, where no correlation between Dielectricity and Electrical conductivity is shown.

- There is a strong correlation (over 0.70) between Dielectrity values in the base layer Port 0, Port 1 and the sub base layer Port 2 and between sub grade layer Port 3 and Port 4. As shown

Air Cooled Blast Furnace Slag & Natural Geological Materials

MASTER’S THESIS

Zana M. Sherif

Analysis of data obtained from Percostations

Differences between data obtained from

Air Cooled Blast Furnace Slag & Natural Geological Materials

Analysis of data obtained from Percostations Differences between data obtained from

Air Cooled Blast Furnace Slag & Natural Geological Materials

Zana M. Sherif

Foreword

Abstract

Table of Contents

1. Introduction

2. Air Cooled Blast Furnace Slag (Hyttsten)

3. Percostations

4. The test road

5. Data analysis

6. Results