llHlHIHlI H I I ('3 34—; '. —80 N (/3204:- I DETERMINATION OF THE THERMAL CONDUCTIVITY FOR SELECTED ROAD SURFACING MATERIALS USING LIMESTONE AGGREGATE Thesis for the Degree of B. S. MICHIGAN STATE COLLEGE Rose Mary Carroll 1949 THESIS Deter-inution of tum Thor-31 Conductivity for Belooted Road Surfacing Hateriala Uaing Limestone Aggregate A Thesis sunnittod to The Faculty of MICHIGAN STATE COLLEGE or AGRICULTURE.AND APPLIED SCIENCE by Rose nary Eggyoll A Candidate for the Degree of Bachelor of Scionoo Juno 19h9 THESIS m ACKNOWLEDGEMENT The author would like to thant.flru Lawrence D. Childs for his advice and assistance in the entire project; Hr.‘Warren T. Edinborough for his advice and assistance in the operation of all the equipment, and Mr. James T. Anderson for his advice and assistance in the Operation of the guarded hot plate. {2.16933 that! OF CONTENTS Introduction . . . Factors Involved in the Heasurement ct Therlal Conductivity . . The Guarded Hot Plate Test Procedure . . Sources of Error and Their Significance Suggestions for Improvement and Further Study Bibliography . . . C O O O O O O Page 13 17 19 TABLE OF ILLUSTRATIONS Guarded Hot Plate . . . . . Hot Plat‘ O O O I O O O O 0 Concrete Sample and Hold Used . . . A. C. Power Circuit . . . . Thermocouple Leads a Direct Current 0 I 0 Circuit Guarded Hot Plate Apparatus . . . . . Sample Data Sheet . . . . . Testnfltfloo...o... Variation in Thermal Conductivity with Temperature . . . 10 11 12 1b 15 l6 Determination of the Thermal Conductivity for Selected Road Surfacing Materials Using Limestone Aggregate The removal of ice and snow from our highways has been a problem without a satisfactory solution. The practice of plowing requires much heavy machinery and is slow, causing many needless delays and the use of chloride, while effective, is harmful to both the high- way and the automobiles that travel over it. These facts have led to the formation of a new theory that has overcome these disadvantages. This theory is the heating of the pavement slab. However, data fro. experimental sections where electrical heating elements have been installed in the road surfacing materials have shown that the operating costs aust be given consideration. For this reason. experiments on surfacing materials for the purpose of determining their coefficients of theraal conductivity have been instigated. This study is a description of the aethod used in finding these thermal constants for three surfacing materials which contained limestone aggregate. The experiaents were as follows: Three aaterials were used with tests being run at two mean temperatures for each, and a fairly flat response curve was deter-ined. Between the extremes of 140° F. and 100° F., the conduc- -1- tivity of bituminous concrete capping material was about 9, standard portland cement averaged about 5.5. and air entrained cement gave values near 5.7. A table showing the exact figures will be found in the section giving the results. FACTORS INVOLVED IN THE MEASUREMENT HERM CON U “The thermal conductivity of a homogenous aaterial is the rate of heat flow, under steady conditions. through unit area, per unit temperature gradient in the direction perpendicular to the area‘.1 ~ In the English system of units. the coefficient of thermal conductivity’t is espressed as B.T.U. per hour, per square foot with.a temperature gradient of one degree Fahrenheit per inch thickness. The lower the coefficient. the greater the insulating value of the material. Mathe- matically, I is expressed as follows: I =n:97?¥:?27_. where Q equals rate of heat flow in B.T.U. per hour. A equals area normal to direction of flow in square feet. :1 equals Fahrenheit temperature of hot surface. :2 equals Fahrenheit temperature of cold surface. L equals specimen thickness in inches. 1. American Society for Testing'Materials. Book of A.S.T.M. Standards, l9h2. p. 1283. -2. Keat flow is the movement of energy particles in search of equalibrium. Therefore, in order to keep the flow perpendicular to the surface of contact, the space around the sides of the specimen must be kept at a tem- perature equal to that of the specimen. There must be a way of measuring the exact heat input and of accurately determining the temperature on the hot'and cold sides of the specimen. THE GHABDED HOT PLATE When determining the thermal conductivity of a relatively poor conductor. it has been found that it is better to use a relatively large area of contact surface in proportion to the distance that the heat must travel through the specimen. To get this large contact surface. a hot plate is used. To insure the flow is normal to the surface, a guard ring is introduced. The guard ring is essentially a second hot plate forming a ring around the plate used for the heat source. This square ring is completely independent of the central element and is separated from it by an air gap one-eighth of an inch wide to prevent any heat flow between the two plates. The central plate current is accurately metered so that the power input can be determined. However. this is unnecessary in the guard ring because its only require~ ment is that the temperature be kept the same as that of ~3- the central plate. A sketch of the hot plate apparatus is given in Figure 1. The profile view shows the arrangement of the two specimens with respect to the hot and cold plates, and also the thermal couple locations. The cold plates are hollow with two connections so that water may be circulated through them to carry away the heat. The whole apparatus is placed in a large box. with heavy cork lining to minimize the affect of room temperature. A differential thermocouple was placed between the hot plate and the guard rind to measure any temperature variation between them. Thermocouples were also put on each of the four contact surfaces. oh- THERMOCOUPLE L E A 03 m L '1 r— ” ' ”Mb m M; 2 t—QZ '3: .l m < w .J o. : a-J z a 3 U .. a 3 0 g 0 (L 0 U 0 I w u P.‘ _..l *‘1 L_ L L GUARDED HOT PLATE GUARD HEATER I CENTRAL HEATER HOT PL ATE Fig. l. -5... T PROCEDURE Samples were molded from bituminous capping material. standard portland cement concrete and air entrained con- crete. The coarse aggregate used in each case was crushed limestone. The bituminous capping mix contained 5.5 per cent bitumen, 5.5 per cent fine aggregate passing a #200 sieve, and 55 per cent coarse aggregate retained in e #10 sieve. The remaining aggregate‘was between these lilits. The mold used consisted of two sections 12 inches by 12 inches and 1 inch thick. This was made of steel and was essen- tially a heavy plate with square bars which were bolted on before it was used each time. This was pro-heated and the bituminous mix forced in at a temperature or about 3000 F. and compacted to a density of about 160 pounds per cubic foot. The concrete was mixed to mold a batch of one-sixth of a cubic foot. The proportions were 3.83 pounds of cement. 5.05 pounds of 2 N3 silica sand. 15.20 pounds of 26A limestone, and 2.50 pounds of water. The same nix prOportions were used for both the standard and the air entrained concrete. A photograph of the-nold and two concrete specimens is shown in Figure 2. -6- Pig. 2. concrete simple and Hold Used. -7... The asphalt samples were ready for use as soon as they had cooled sufficiently but the concrete was allowed to cure for seven days and then placed in the oven for three days to dry out before the test could be run. They were all carefully weighed to determine their density when ready to be tested. Before the test was run, blotting paper was placed on the contact surfaces of the plates to insure good contact between the specimen and the plate, and to keep the temp perature of the thermocouples at the specimen temperature rather than at the plate temperature. The samples were then put in the box between the hot and cold plates and clamped firmly into place. The box was closed and the water pump was started, thereby circulating the coolant through the cold plates. The current to the hot plate and ring was turned on and the power input regulated to get a constant heat flow. After about an hour. it was possible to check the differential thernocouple‘between the plate and the guard ring. The ring power was then adjusted until the tespers- ture was the same as the hot plate. This adjustment was quite critical and, consequently, it was usually several hours before the differential thermocouple between the hot plate and guard ring indicated no variation in tempera- ture between then. In order to conform to the standards as much as possible, the apparatus was held in balanced -8— condition for four hours before any readings*were recorded. Usually, this balance period was accomplished by allowing the unit to remain on all night. A schematic arrangement of the power circuit, showing the separate rheostats for guard and plate adjustments is shown in Figure 3. Temperatures were measured by using copper constantan with a reference Junction at‘32o F. At the start of a record period, an ice bath was prepared and allowed to stabilise. Readings were then taken at about fifty-minute intervals until a fairly level five-hour average was ob- tained. A diagram of the thermocouple circuit is shown in Figure h. The two mean temperatures were determined by running two tests on each speciaen. The first with the water in the cooling system at room temperature and the second.with the water near freezing. The reservoir for the coolant for the first test was a fiftybfive gallon drun, and for the low temperature test, a mechanical refrigeration unit. A view of the complete apparatus with the specimens placed between the plates Just before the box was closed, is shown in Figure 5. CONSTA NT VOLTAGE REGULATOR POWER RHEOSTAT CENTRAL RHEOSTAT ’VVVV\ IIO A.C. ' t GUARD RHEOSTATS GUARD COIL h AMMETER CENTRAL con. l l W u VOLTMETEa A.C. POWER CIRCUIT Fig. 3. -1 0.. .EDUEU Fzmmmau POM >¢uhk- . .— 2 1m '- U 3 0 Z O U .1 ( 6.0 c GQ’ ENTRMN o u I - r. w \> 5.0 40 5O 60 7O 80 9O IOO HO ' MEAN TEMPERATURE-DEGREES F VARIATION 2'22 THERMAL CONDUCTIVITY 2222/2 TEMPERATURE Fig. 8. -16- readings made it possible to detect these errors in the temperature circuit but a short or a broken lead in the differential thermocouple circuit could not be readily discovered. Part of the errors came from the test specimens . themselves. Limestone aggregate is a slightly better conductor than standard concrete. Each material tested became a better insulator at a higher temperature. The coefficient of thermal conductivity for these materials for use at a low temperature, as in ice melting installations are: - Bituminous concrete - 9.1 Air Entrained concrete - 5.9 Standard concrete - 5.8 gUGGESTIONS FOR IMPROVEHENT.AND FURTHER STUQ;_ Too much faith cannot be put in the results of the test because only two runs were made on each specimen, making it impossible to check the curve. There is no Way of checking the accuracy of the equipment. The tempera- ture varied during the day, making the temperature of the cold plate vary over a period of time and the guarded hot plate itself is too small to accurately determine the thermal conductivity of specimens which have as high a value as those tested. Three or more trials should be run on each specimen -17- tc check the curve and if they did show the slope to be constant then, regardless of the accuracy of the equip- ment. the tests would fulfill their purpose in that they would give a definite visible comparison of the actual performance of different road building materials. ~16- BIBLIOERAPHY 1. American Society for Testing Materials, Book a; A.S.T.M. Standarde. 19%2, pp. 1232-1290. 2. Anderson, James T., h e i n r Guarded R . Hg; Plate {0; Tegtin th Th, 8 C nd ct of Homorenou Ma eriala; ihesie for Degree of H. 8., Michigan State College, 19h8. ’ 3. Hanenann, Brick; and Slack, Edgar P., Ph ic . D. Van Noetrand 00., 1935, pp. 318, pp. 89. -19, E R I: I I J MT! 03082 6840 [Mir 1 MI £HIWHEIIHWIHEJ II a‘lHH N “:1 3 1 93 W