s—v‘y . ) q.1 H ‘ 5’1: 4 got; if E. 5 THE THERE‘ML connenm {JF EXPANSiON 3 E ‘ UF LUMNITE swam WJRTARS THESIS FOR DEGREE OF 23. s. E i z EN CIVIL ENGINEERING '. a. A. GEFFELS ; i 1 9 :2 i“ r WW" - UH 58:5 009‘ Fliqll. .9. i 1.51mi THE THERMAL COVFF‘ICI‘TNT 0? EXPANSION OF LUMNIT‘L‘ G‘TWCNT MCRTARS A Thesis Submitted to the Faculty of THE" MICHIGAN STATE COLLEGE of AGRICULTURE e APPLIED RCIWCE I Q" gIrVPS c. i." guns Candidate for the Degree of Bachelor of 801 enoe June 1926 id-au" ' -ir-' mall. jcus‘ ‘ UHESJF ~ ‘ Qay.‘ In reinforced concrete construction it is important that the thermal coefficient of eXpansion of the concrete is ab- out the same as that of the reinforcing steel. Otherwise considerable stress may be caused in both the steel and con- crete by a variation in temperature from the temperature at which the concrete set. The nescessity for empansicn Joints is also controlled by the coefficient of expansion. Portland cement mortars and concretes have been tested, and the thermal coefficient of eXpansicn has been found to agree closely with that of steel. However, a new cement, which is likely to become popular, because of its fast sets ting prcperty, is now on the market. This is called Dum- nite cement, and in America is made by the Atlas Lumnite ce- ment Company. The purpose of this thesis is to eXplain the method used, and to give the results of, some tests made to determine the thermal coefficient of expansion of mortars of sand and Lumnite cement. The apparatus used was the optical lever, telescope, and scale used for such work by the Physics Department of the Michigan State College, and a standard and steam jacket made for the purpose. with this apparatus, the coefficient could be determined within 0.0000001 inch, if care was taken. ’ The specimens were heated by steam, since this provided a sure method of uniformly heating them, and determining the temperature. The steam jacket was insulated with asbestos pipe covering, so that the posts of the standard would not be effected by the heat from the steam jacket. QCSJG In order to avoid OXpanllOn thru moisture, the specimens were cured and stored in water. Two specimens were made of each mix tested, each being tested once, and the average co- efficient was taken. All specimens were tested at Bl days. The original intention was to cool the specimens in ice water, getting a wider range of temperature than by starting at room temperature. However, it was found to be difficult to maintain them at that low temperature while putting them in the steam jacket, and getting the first reading. The time needed to get full eXpansion was only from 30 minutes to 30 minutes, and the specimens would warm up appreciably in warm air in a few minutes. The specimens were bars of mortar 38 inches long, and 1 inch in crossection. They were made as nearly of the same consistency as they could be, considering that they 3/“ were made on different days and of 4$££3£22£a21:3:;,ew953”’ normal consistency of Lumnite cement was taken from a bull- etin by Mr. 323R3&IY. of the Civil Engineering Department of the Michigan State College. This value was 25.&% water. The percentage of water recommended by the American Society for Testing Materials for standard Ottawa sand mortar, 1:3 mix, to correspond to a normal consistency of 36%, is 10.8%. A sample of standard mortar,l:3 mix, was made, using 10.8% water. This gave a standard for comparison in making the specimens. The sand used for the specimens was screened thru a 30 mesh sieve, from bank run gravel. The results of the tests, as shown in the table, indicate that the cement is suitable for reinforced concrete work. The coefficient of eXpansion of 933: Lumnite cement is , 0.0000071 as compared to 0.0000078 for Portland Cement. The coefficients for the mortars range about 0.0000060 while the coefficient of eXpansion of reinforcing steel is about 0.0000065. ' The eXpansiOn of the mortars decreases as the percent- age of cement is decreased, until a minimum is reaChed with 1:3 mortar. From then on the exnansion increases. There is no apparent reason for this, since, if the coefficient of . eXpansion of sand is less than for the cement, the exnansiOh of the mortars should decrease as the mixture is made weaker. The only other eXplanations that seem to offer a solution, aside from an improbable variation in the nature of the sand, are variations in the consistancy of the mix, or variations in mixing or compacting of the specimens when molding them. A indication that one of these three is the.proper exclana- tion is the result obtained with a specimen of 1:3 mix, us- ing 15% water, rather than 15.4%. The coefficient found for this specimen was 0.0000060 as compared to 0.0000064, which was found to be the coefficient for the specimens using 15.4% water. Such results show that the coefficient of eXpansion cannot be predicted for any particular mix with accuracy, unless great care is taken to get exactly the same conditions in the mixing and handling, as were had in making the test specimens upon which the prediction is based. In other words, in prac- tics the coefficient of expansion cannot be exactly predicted. Howefsr, these variations do not seem to amount to enough to seriously effect stress ——-—..‘. ..,. .— mixture neat 1:1 1:1 1:2 1:2 1:3 1:3 1:4 1:4 % water 10.8 16.4 16.4 16.5 16.5 15.4 15.4 16.0 16.0 Table temperature scale change reading 135 37.0 148 37.0 148 35.0 138 30.0 139 30.5 145 36.0 143 35.0 145 38.0 145 38.0 is in concrete or steel. coeff. .0000071 .0000065 .0000061 .0000056 .0000057 .0000064 .0000064 .0000068 .0000068 av. coeff. .0000071 .0000063 . 00000565 .0000064 .0000068 amni‘fi siu‘ ‘4 H 'i V A UNIVERSITY LIBRARIES 03 06] 4154 NT“ 93 "lolliffilmflllfifl 3 12