:2: :_.E_:__E_E_E=__:=_:=:_= m 8 2 8 S H T . .I); a; Q. x. :o 3.03 — ; J, I. «I In . . I'll ' l‘ l | Illll ll. »\ .. . 2‘ 5‘ q .‘l a A: v .1 u .. v. n u a .... . u \e .. {.77. $19,. u .. .u . ..... 9. 1.: .v«A ,34 ..,“..‘. .5! ‘1 ..J\...:,a..,v. ., ..i r”. ...‘ . t! . .h 7 A‘ .,.. .7 ...$.. ‘uol . ...i..1o (fit... N ..)§.t.. . . . ‘1... It .... , .. 3. FT 4. f. ,. ... V: .. ‘ 4 .«r:2. . Z W. . x f L ., . ‘C. a» . r ..L.. c . . . .. 5 .133 .\\...xiwp.1uwv»: .1.3§V.1.... .1, ix! .linw w... L.Lfiwfmgfirgmfiwvmumfi.. C. {:. I . . y .. i5? 14.22-11,. EBEQ: . . . / mic ‘ J ‘ ~ .. / K , . I O“ .v . . . ‘ I J.‘ _. ‘IU\ .. , , , J I [kw . v L, ‘ x ., Ix. A , .4. If... , h L 1 r . bx Lu m‘ 0 ~ I e . “ . I L r . ~ " I. ,. . l . k . P , *\ x b 4.4 \ l HRS TESTS OI CORRETE ~ ~ . A Report Bnmtted to the Penalty 0: 11:0an “3100me COLLEGE 3? Avery Judd ”Egan; name Prederiok Barrie .tal" candidatee for the Degree ‘ of BACHELOR or SGIEBCE. June, 1984. THESiS LOPW ”Concrete Engineers Handbook" Ecol and Johnson. "Engineering" Proteeeor P. 0. Lee. "lire recto“ fictional Board or Fire underwritere. "Report on the Fire or the Edison Phonogrcph Worke' lotioncl Fire Protection Association. 94231 "We wish to express our appreciation and thanks to the follow- ing man for their kindness and willing- ness to aid and advise as in the maintenance of this Report." Professor H. K. Vedder, Professor 0. Allen, Associate Professor B. Sangeter, .Mr. Luther Baker. Mr. S. L. Christensen. 7% fizz-MM / I, 7") x- ‘ / .-/ ' 'V I/ J A... U FIRE TESTS OI COHCREEEE. In maintaining this report, "Fire Tests on Concrete” it is the sriters purpose to show both by experiments of their own and experiments carried on by the National Board of Fire Underwriters and other organisations, the effect of fire upon concrete, both plain and reinforced. Plain and reinforced concrete structures are very common in these days, and buildings made of this material are frequently referred to as fire-proof. In the United States as sell as in Europe and Canada, disastrous fires have, however, frequently proved that providing a fire rages for a considerable length of time, as is often the ease in large warehouses, these structures fail badly and in neny cases it has been necessary to dismantle the building completely. By experissnts of our own and experiments conducted by the Fire Underwriters and other organisations, so have been able to arrive at sons very interesting and definite conclusions. It is our aim to show to Just what extent concrete can be considered as fire- resisting. In this report will he considered the action and effect of fire upon plain and reinforced concrete; columns, sells, beams, slabs, and partions. In considering the effect of fire upon reinforced concrete the problea can be stated under tso headings: 1. Is it possible to lake concrete which sill retain its strength during and after exposure to high t-peratures liable to occur in a building fire? 2. Is it possible to prevent the steel from reaching such a taperature that its strength is reduced to or tales, that required to carry the load occurring in the sue building? Concrete essentially consists of frapents of stone held together by a net sort of sorter. This sorter in itself is a fine aggregation consisting of fine grains of sand "stuck" together by Portland Cseent. An erasination of concrete shows one peculiarity very strikingly - it is porous. The Voids vary in size free: these easily visible to the naked eye, to a mass of fine channels and cavities of uicrosoOpie dimensions. The node of failure of concrete may be as folloss: l. The concrete can be considered to consist of a network of cement holding together stones of various sizes; if the finer aggregation I1 of mortar is caused to fail the whole a" will be disintegrated quite independently of any effect the coarse aggregate may have. 2. In such a network structure if all the stones are covered with cunt no increase in total volume can take place unless the scent itself expands. mm expansion takes place the stresses will be produced in the coating canent and in the aggregate, unless the expansion of the assent is exactly equal to that of the aggregate. whose stresses may be sufficient to cause cracking of the cusnt, or when cooling takes place may lead to separation between adjacent bomdaries. This being the case concrete nay fail in high taperatures due to the different coefficient of expansion of sand and cement or it may fail by the scent itself breaking down caused by sons innate property of the essent, which would take place whether the sand were present or not. he coefficient of expansion of quarts, which is the chief constituent of all sends used in practice Ponclav?‘ (JONG‘ 473’ Expo? J/o/V Cfiner Imam—5:32 ave/v Co” Turner/02v U/ xao 200 400 600 7‘MJR077/RE actor‘s-r: CfiA/‘T‘EGRHDI. has been fairly accurately determined and expands at a quit uniform rate. Cement is entirely different. The results of a number of tests show that up to 100°C, cement has a fairly steady increase in expansion, but at 100's, a very large contraction is started, this contraction continues until a temperature of 491°C, was reached. it this ties the cement has contracted am it is such manor than its original size. After 491°C is passed expansion sentences, and takes place at a rate less than that occurring during the expansion previous to 100°C. Figure (1) gives a graph showing the relation between tauporatire and expansion of scent. - Another thing to take into consideration is the effect of heat on hydrated and hardened portlsnd cement. It was found that water was given off at a fairly fast rate up to 110°C. Pros there on water continued to be given off but not as rapidly as before. There can be little doubt that the contraction obtained in the eXperinents and described above, as designated to ascertain the value of the coefficimt of expansion, is due to this dissociation of water from the hydrated cement. :t is micrograpl magniiied 100 diameters. Figure 2 is a photo-nicrograph of a staple of set concrete which has been aagnified 100 disasters. The method of etching causes the softer portions (dark in the picture) to be rubbed away and leave the harder (white in the picture) standing out froze the surface. By further examination of these harder portions, it can be shown that they are unchanged particles of Portland cement clinker, that is, grains which have never been hydrated by the airing of the concrete. In this case we really have hydrated scent surrounding particles of inert and unchanged clinker, the presence of which is partially due to the coarsness of grinding. Though these particles are fine enough to pass a 180 x 180 standard seive. is the taperature' rises lore expansion of the unhydrated frapents will take place causing hydrated cement to break down. This action will continue until all the unchanged clinker is free to move and expand when the specimen on further heating nay expand or contract, depending upon the relative novments of the hydrated and unchanged portions. J Fru extensive experiments carried on in England we learn that when concrete is raised above 100' I, then contraction will occur. On the other hand steel ilbedded in the concrete will continue to expand as the t-perature rises. Uhen this happens the adhesion between the steel and concrete auet break down either by a complete sliding of the steel through the concrete or the concrete nust crack and leave the steel in this nanner. In practice it is usually the latter which occurs. With the snount of concrete covering ordinarily allowed in design it is scaewhat doubtful whether the question of heat conductivity of the concrete is of primary ispertance. This spelling action will, in many cases, occur long before the t-perature of the steel can have risen to the degree which is dangerous. This spelling is caused by the sudden and intense heat which produces a rapid expansion of the concrete near the surface of the column at right angles to each other, subjecting each corner to stress fron two directions tending to force the corner off in the line of the diagonal. it the same time the surface concrete, especially at the corners where heated on two faces tended to expand lengthwise and thereby assune an undue proportion of the colmn load with the result that shearing stresses are produced between the highly heated corners and the colder interior portions of the column. This action occasioned a buckling effect of the concrete at right angles to the length of the calm. The combination of these forces will produce tmsile and shearing stresses in the diagonal planes across the corners, which results in splitting than off lengthise. This spelling action will be seen to be the nest serious at the underside of beans and the sides of colms. If a concrete that will not spell can be discovered the conductivity will be the only, and not by any neans the most serious menace. These experisente were performed within a few hours after the specisan had been taken from the fire. But by accident one piece was not tested for over a week. It was found to have lost strength quite out of proportion to the snount of heating. when a block was taken fra- the furnace no cracks on the surface were noticed but Ihll exposed to the atleosphere of u:- laboratory, nunerous cracks occur and the block may even crusble. ' The explanation of tie phenuencn can be found in the fact that one of the chief products of hydration of Portland ceaent is oalcuia hydrate. This - dissociates into quck line and water at about 400°C. 0. (on), .... cao . ago 0—- It must be realised that this is accompanied by a contraction of the concrete. Concrete is porous and the air getting in will carry moisture with it. This moisture will hydrate the quick line again and the product of hydration (Ga (mm occupies a considerably greater voluns and base causes the concrete to crack, split, and ultimately crumble. The importance of this "after effect of fire" cannot be over estimated. When the question has to be faced as to whether the building is after- wards safe, and if not, how much of it should be duolished and rebuilt, and the answer would appear to be not very maul-aging. In view of the large nuaber of buildings already erected of reinforced concrete, the problas cannot safely be left at this point, but in the Opinion of the writers the next work should be done on full sised structures that have undergone these conditions. In view of this fact we will first consider colmns. While column form the acct important slnent in the strength of a building, few representive tests have been made to detsrnine their abiliw to support load whu exposed to fire. {the purpose of the following experinents, run by the National Board of Underwriters, on full siaed columns, was to ascertain l. The ultilats resistance against fire of protected and unprotected columns as used in the interior of buildings. 2. Their resistance against impact and sudden cooling from hoes streams when in a highly heated condition. _The fire test series includes:- 1. feats of representive types of unprotected structural steel, cast iron, concrete-filled pipe, and timber columns; 2. 'i'ests where in the motel was partly protected by filling the reentrant portions. or interior of colunns with concrete; 5. Tests wherein the load carrying elusnts of the colunn were protected by a 2 inch or 4 inch thickness of concrete, hollow clay tile, clay brick, gypsum block, and also single or double layer of notal lathe and plaster; 4. Reinforced concrete calms with 2 inch integral concrete protection Although our chief interest is in tho effect pf fire upon concrete these other tests will give a fair and Just comparison and at the same time show the advantages or disadvantages of concrete against some of the other materials. at the same tine the full action of the fire tests upon the concrete will be shown. the test columns were designed for a carting load of approximately 100,000 lbs., as calculated according to accepted formulas. The load was seintained constant on the calm during the test, the efficiency of the column or its covering being determined by the length of tine it withstood the combined load and fire exposure. the latter was produced by placing the ealusn in the chamber of a gs fired furnace whose t-perature rise was regulated to conform with a predeterained time-tapereture relation. leasurseente were taken of the tnperature of the furnace and test column and of the deformation of the latter due to his load and heat. In the fire and water tests the calm was loaded and exposed to fire for a predetermined tine, at the end of which the furnace doors were opened and a hose stress applied to the heated calm, he duration of the application and pressure at the uncle varying with me length of tine the corresponding type of column withstood the regular fire tests. All calunns tested were of 12 ft. 6 inch effective length with an additional 3 ft. to take up the load and transmit it to the calms. is our chief interest lies in the concrete columns it will suffice to say that the other colunns tested were all determined as to size and loading by well known fornulas and a ccnpariscn lads between the final results of the‘ concrete protected columns against those not protected. so will discuss the calms in the following manner: a. Columns Protected by Concrete. Under this test Rolled H, 2-bar and plate, plate and angle, plate and channel, latticed channel, lobesn and channel, starred angle, latticcd angle, round cast iron, columns were used protected by Concrete 2 inches to 4 inches thick. Six combinations of fine and coarse concrete aggregates, as used in building construction in four large industrial centers, were used; namely:- 1. Rochpart granite with Plum Island sand for Boston, Licssaahusetta district; 2. Chicago lilsstons with Fox River sand, and Joliet gravel with Jaliot sand for the Chicago district; 3. Cleveland sandstone with Pelee Island sand for Cleveland, Ohio, district; 4. low for): trap rack with Long Island send, and hard coal cinders with Long Island send for the new York, H. I. district. Portland cement was used throughout the tests. The proportions of the mixture used were 138:4 and 1:3:5 for the stone and gravel concrete and for the cinder concrete l;l%;d% and l;2:5. The cinders were used unsoreened except that pieces larger than 1 inch were crushed to usller sise. Ties consisting of Na. 5 (3.8: 8. gauge) bright basic steel wires were wound spirally around the structural section on vertical pitch of 8 inches. The porpartioning of aggregates were based on volume parts of the materials except that the Portland agent was measured in the original package. The sand and stone were measured in deep steep wheel-barrows, the value of each being determined by a tnplet of the required shape. ill concrete was mixed in a motor driven concrete mixer. Everything was conducted as nearly as possible to the actual field work. The sand and coarse aggregate was all tested according to the required tests for each, and was then parparticned in the right amount for the different mixes. The concrete specimens were cylinders 8 inches by 16 inches. These cylinders were tested, for each column, in the usual manner. The loading apparatus was a special hydro-pnepatie run which was designed to maintain a constant load during the test. The calms usre moved to their place in the furnace by a carriage built especially for this purpose. The furnace was heated by means of four primary blast burners arranged to discharge in an inclined direction upward and toward the adJacent corner. The burners are supplied with gas. ill teqerature measur-ents were made by the thermo- electris method. The calms were measured for three kinds of deformation: 1. The unit compression and expansion over a definite sense 10118“. 2. The total depression or expansion of the column measured at a point above its heated portion. 3. The lateral center deflection. In case the column withstood the 8 hour fire test it was i-ediately loaded to failure under full fire exposure. In the fire and water test series, the protected structural steel, the two unprotected cast, and the three reinforced concrete columns were loaded to failure after they had cooled. The duration of the fire periods varied fru 88} linutes to 1 hour, and that of the subsequent water application, fron l to 5 minutes. The length of the maximu- fire period was determined by the time within which water is generally applied in building fires, which is estimated at one hour. In applying the hose stream, the ncssle was moved back and forth on one side of the furnace and maintained at a constant distance from the column, the water being applied in succession over the full height on three sides. I. Reinforced concrete columns. Tests were run on tree kinds of columns: 1. Square vertically reinforced, 2. Round vertically reinforced, 3. Round hooped reinforced. They were all made of a 1:2:d mix and the nterial, apparatus, and manner of testing was the same as in the proceeding Section A. It appears free the tests of concrete applied as a protective covering or filling to steel or cast iron coluIns, that the concrete retards the tasperature rise in the metal when the calm is exposed to fire and further retards the failure by carrying portions of the column load proportionate to its relative area and rigidity as compared with the metal. The protections were applied as square or round coverings, generally 2 inch and 4 inch in thickness, measured frms the surface of the ca vering to the metal. The time of failure in the fire tests, varied from 1 hour, 47 minutes, to 7 hours, 57 minutes for the 2 inch protections, and fron 3 hours, 47 minutes to over 8 hours for the 4 inch protections. lo evidence was developed that variation in the strength of the concrete of the acne aggregate and pcrporticn of nixture had an appreciable influence on the results of fire tests of concrete protections. This was due to the large change in nechmical properties produced by the heat. Concrete as lads with different aggregates preserves strength to different degrees on exposure to fire. With a given thickness or sise of covering the main cause of variation in results was the difference in fire resisting properties of concrete made with different aggregates. In this particular the concrete can be placed in three groups: i'hat giving the most unfavorable results was the concrete ads with llerinse River sand and gravel. This was due to the fact that this sand and gravel consisted almost wholly of quarts and chert grains and pebbles, the gravel having a particmrly high chart content. Both minerals are found in silica (Si 02), the quarts being crystalline and anhydrous, and the chart amorphous with a variable snount of water in cheaical combination. m being heated part of the combined water in chert is liberated and the consequent vaporisation disrupts the pebbles. Other causes of disruption of concrete made with siliceous aggregates are abrupt volumn.changes. The .columms.mads of this gravel with a high silica content gave for the covered reinforced concrete, from one hour fire resistance to two and one half hour fire resistance depending on the thickness of covering. The reinforced columns also showed that this kind of aggregate was very poor and only gave a five hour fire resistance tect. The middle group includes concrete made with trap rock, granite, sandstone and hard coal cinder. In tests with trap rock and cinder concrete a small amount of cracking occurred and during the last part of the fire period, no spelling of am note occurred before failure. In the granite concrete cracks develOped earlier in the test and spelling took place during the last 30 minutes of the test period. In the test of the sandstone concrete, cracking and spalling began.in the first 30 minutes and continued for an hour, after which there was little apparent change before failure. Fusion.of the trap rock concrete occurred where the test extended‘beycnd seven hours, the concrete being affected to a depth of about 1% inches. Flowing of concrete due to fusion, while not general, occasionally formed pockets up to 2 inches in.depth. The third group comprises protections of Chicago limestone concrete, and Joliet gravel concrete. The composition of this gravel is similar to that of Chicago limestone and the tests compare quite closely. Very little cracking resulted on.exposure to fire and their host insulating value was increased.by the change of the calcium.and magnesium carbonate to the corresponding oxides. This process retarded the flow of heat through the region of change and left material of good insulating qualities. Immediately after test the surface of the concrete was fins, but after a few weeks exposure the hydration of the oxides caused slacking and crumbling of the calcined material. From a comparison.of the thickness of protections it shows that a four inch protection.is much better than a two inch protection, although there was only a difference of a few hours and.the good two inch coverings all withstood the 8 hour test. Concrete as made with different aggregates preserves strength to different degrees on exposure to fire. This had a decided influence on results, the longer test periods and particularly the longer intervals between maximum expansion and failure of the limestone concrete and Juliet gravel concrete can be attributed in a great part to this cause. There seems to be little difference as to the shape of concrete, coverings or columns, as concrete made with a highly siliceous composition makes the other defects small in comparison to that of the aggregate. In covered structural steel columns it is much better to have a wire netting around the column as it tends to hold the cmcrcte to the steel. In the case of the water tests the covering was carried away frms the unprotected columns while those where wire was used the concrete held fairly well. In regard to the reinforced concrete columns the experiments show the following results: ,, The limestme concrete calms all withstood the 8 hour fire test and while hot sustained loads exceed- ing twice the load applied in the 8 hour period. The two vertically reinforced trap rock columns failed after 7 hours, 225 minutes, and '1 hours, 5'7 minutes, respectively, and the hooped column withstood the 8 hour fire test and failed under a load about 25 percent greater than the load, sustained during the fire test. A 2 inch thickness of concrete next to the surface was assumed as covering in all cases and not included in the area used in computing working loads. The difference in results within the group can be attributed to concrete aggregate, the other incidental factors being comparable to, or favoring the tests giving the lower results. The trap rock concrete fused and fluxed at some points to a depth of one inch, which undoubtedly affected the time of failure to some extent. The results obtained with the concrete of both aggregates show a high degree of fire resistance. lo effects due to shape of calm or farm of reinforcement were evident, differences in results being within the limits of incidental variations in test columns and conditions. No line of cleavage outside of the wire reinforcwnsnt was found after test in the hooped column of limestone concrete, except in the immediate region of failure, where it was apparently induced by the strains that developed when the column failed. In the case of the corresponding trap rock concrete column, more evidence indicating separation of the outer protection from the care at the line of the reinforcement was found, effects in part which may have been caused by the fire exposure. One length of the hooped reinforced concrete columns about three feet long was out outside of the failure region in the fire test and subsequently tested in compression. The limestone concrete specimen sustained a total load of 517,000 lbs., as against 243,000 lbs., immediately following the fire test, and the trap rock cmcrete specimen, 342,000 lb. compared with 165,000 lbs., at the end of the fire test. The greater part of this variation in strength can be attributed to recovery in strength of concrete and reinforcement. f The concrete of the columns subjected to fire and water tests was placed in three sections to permit using two or three kinds in each column. In the case of the square vertically reinforced calm, the water carried away the concrete at the corners outside of the bars and pitted the * concrete on the most exposed face to depths of from 1/8 inch to 1 inch for the limestone concrete and to a depth of 2 inches for the Meramec River gravel concrete. In the round vertically reinforced column, the limestone concrete was pitted to a depth of 1 inch and some of the concrete in the upper portion of the Joliet gravel concrete section was carried away. That consisting of heramec River gravel concrete, the outer concrete was stripped off by the water, exposing the reinforcing bars on two sides. In this as in the proceeding large cracks had formed in the concrete during the fire period. In the fire and water test of the hooped reinforced concrete column, the water stripped the Marines River gravel concrete and the granite concrete ‘ r . .‘F’ Ag" '.‘_ - J. H I '._."__ -. . I'! ' .a \ g (0 . ' .. . Q g . er-.- 5 I .O‘ .' _)«3 I -&' J ‘ . r! ,1 IL.’ a ‘ ‘ .'{. ' g , f r." ‘ ‘ ' . . ‘ (_ t v NC I ‘ . '\., .r,,4 :‘ ‘,.'s -' “‘3? Jo. V , ’ 3’ .‘ 1'5'.‘4.,7 : 2 . ,' o l (, ' \r '. ‘ ‘,_ w. v ‘I .. '1‘ u ‘l ‘\ . ‘ # ’ I _ ‘ .I-si. Note the concrete hanging from the under side of the girder in small stalactiyes. from the wire reinforcanent on three sides during the first 16 seconds of the water period. Spelling cf the concrete had exposed portions of the reinforcement during the fire tests. Further application of water caused stripping to the reinforcement in the upper section of trap rock concrete and increased the effects in the lower sections. Irma all of these experiments a great many things are determined but each large fire brings some new unforseen thing to ones attention. Perhaps one of the worst fires occurring in a concrete building was that which destroyed the "Edison Phonograph Works" in West Orange, New Jersey, in 1914. Perhaps one of the most interesting things learned from this fire was the fusing of concrete in the basnent of the Wax house. The ceiling, beams, girders, and columns supporting them in this lowest story show remarkable appearance of fused concrete (see figure). Small stalactites of concrete slag hug down frm the ceiling and large bunches of it achere to the columns where it has run down and hardened. The lower part of these three beams has wasted away, exposing the reinforcment which has melted or burned out, causing failure of the beam. The under side of the floor slab adjoining these beams has also wasted away, exposing the metal reinforcement parts of which has melted and hangs down in tapering rode, the area of which has been reduced in some cases to about one eighth of its original sise. The columns, in addition to being fused on the surface, have also spelled. This would indicate a temperature of 2500?, or more at this location. The action of the columns was not as satisfactory as that of the other concrete manbers. Host of the columns were square and reinforced by twisted bars located at the four corners two to four inches from the surface. The sudden and intense heat produced rapid expansion of the surface concrete, resulting in severe internal stresses. The corner reinforcement bars also tended to produce a plane of weakness in the concrete, many of the comers splitting off along the line of these two bars. In some cases the bars may have expanded sufficiently to have aided the splitting action, but the indications are that most of these corner failures originally occurred outside of the bars, at a point where the stresses more- than equalled the tensile strength of the concrete in the diagonal plane of failure. With the corners removed, the reinforcing bars had little protection from the heat, consequently expanded rapidly forcing themselves out of the columns, and carrying with them any attached surface concrete. The coefficient of expansion of the concrete is approximately the same as for steel, and as all the heat reaching the steel bars would have to pass through the concrete covering, it is probable that the former would always be a little cooler than the latter. 'This assummicn is strengthened by the fact that one side of each bar was in contact with the cool interior concrete. Furthermore the thumal conductivity of concrete is low. I In spite of the disintegration of the surface concrete, it is probable that if the bars had been prcperly tied, many of thu would have stayed in place and the column injuries would have been less severe. In most cases it was the wall columns that failed. The greater injury to wall columns is believed to be due to two causes. first, to the fact that while subjected to intense heat on the room side, the opposite side exposed to the outer air was kept comparatively cool, and was in some cases subjected to cold water, from hose streams; Second, owing to the columns being held rigidly in position vertically from the floor to the top of the panel walls, they were less able to resist the expansion of the building as a whole, which naturally resulted from the attack of fire. It is probable that reinforced, and that they were free to bend from.flocr to ceiling in the oirection in which the building expanded. This theory of the bending of the column due to the expansion.cf the building is further substantiated by the existence in all cases cf’a V-shaped crack.bs- tween the corner column and the adjacnt wall panel, start- ing at the floor and widening upward. While these various theories and descriptions may explain what actually happened in the sequence of events proceeding the.failure of different structural parts of these buildings, they do not Justify the general conditions which produced the results which confront us, and unless the experience thus gained will insure that future specifications shall be drafted to prevent a repetition of such a disaster under similar conditions, it will indicate that either the lesson has not been properly learned, or that the design and censtructien of these buildings is not suited for the purpose employed. Whether any other system of construction would have given better results under the same conditions, is problematical. The lesson is being carefully studied by many competent persons. The knowledge gained, will doubtless aid in eliminating unwise practices in reinforced concrete construction. It has needed an expensive lesson of this_kind to demonstrate the strict necessity for xhanges in design, and methods of construction to meet such conditions. Reinforced concrete buildings can doubtless be built which would withstand such a fire satisfactorily, but no type of construction should be left to meet such an attack without the assistance of any of the standard fire resistive measures which should be a part of every first class building. It appears from an examination of these tests that structural steel columns protected'by 2 inches of concrete withstood the fire tests fro-.1 hour, 45 minutes, to 7 hours, 67 minutes. Structural steel columns protected by’e inches of concrete withstood the fire tests from 3 hours, 40 minutes, to 8 hours, 30 minutes, and these required.an.additicnal load before failure occurred. Reinforced concrete columns withstood the fire tests from 7 hours, 85 minutes, to 9 hours, these failed under an additional load. The other forms of columns did not reach any such limits - most of them failing under a 5 hour fire tests. Hence, it can be said that with a proper aggregate, good reinforcing, and everything carried on in the proper way, concrete affords very good fire protection.and unless an unusually large and furious fire occurs there is no better fire resisting substance than concrete. The purpose of our experiments is to ascertain to what extent fire will effect concrete. To do this we used material purchased on the retail market so that we could perform our experiments with material as nearly like that used in regular construction as possible. We also wished to determine the difference in effect, if any, upon cement manufactured from marl and that manufactured from.lhmestone. The materials used were those that were purchased in Lansing, Michigan, The cement used was known.by the trade name of new Aetna (a marl cement) and Petoskey cement (a limestone cement). The coarge aggregate was washed Mount Hope gravel. This gravel is secured.from the.lount Hope gravel pit Just outside of Lansing, and is the same as that used in the construction of four large buildings on the Michigan agricultural College Campus at the present time. Standard Ottawa sand was also used. the tests were run in three series, namely, tests on briquettes, cylinders and beams. BRI QUETTES . The normal consistency of the cement was first found by the Vicat needle consistency test. This test was repsated four times and an average taken. The consistency of the Petoskey cement was found to be 27, and that of new Aetna, 27.25. The ’ briquettes were made as recommended by H001 and Johnson, a standard method. The cement is mixed for 1.5 minutes and then pressed into the moulds firmly with the thtmbs and smoothed with a trowel. The briquettes were tested on a standard testing machine in the cwent laboratory of the Michigan Agricultural College. THOSE MADE FROM.MARL CEMENT. NEAT CEMENT. lO - tested after being subjected to fire, 10 - before being subjected to fire, 1 - 3 Standard sand. 10 - tested after being subjected to fire, 10 - tested before being subjected to fire. THOSE MADE PROM LIESTOKE CEMEHT. NEAT CEMENT. lO - tested before being subjected to fire, 10 - tested after being subjected to fire. 1 - 3 Standard Sand. tested before being subjected to fire, tested after being subjected to fire. 10 10 These samples were placed in water after 20 hours and remained there for 24 days, at which.time they were taken from the water and allowed to dry for 6 days before testing. CYLINDERS. The cylinders were made of the above mentioned gravel and were of a l + 2 - 4 mix. They were 6 inches in diameter and 12 inches high. The forms were rmoved after 24 hours and the specimen allowed to stand in the air for 87 days. They were tested for compression.cnly cn.a Standard compression machine in the Strength of materials Laboratory at the Michigan Agricultural College. ' MADE OF EARL CEMENT. g. I tested before being subjected to fire, h I tested after being subjected to fire. MADE PROM LIMESTONE CEMENT. .p. I tested before being subjected to fire, g. I tested after being subjected to fire. BEAMS . The beams were made of the same material and mix as me cylinders were made of. They were limited in.size by the furnace in which they were to be heated. Pour specimens were made, two being of marl cement and two of limestone cement. These beams were tested on Standard testing machines in the Strength of Haterials Laboratory of the Hichigan.Agriou1tura1 College. DESIGN OF BEAM. Assume 2000 lb. concrete. fa - 52000 lbs. 1'9 . 32000 to 2000 O 16 K . 1 . e652 1 e 16 a - 1 - 3/8 x .552 - .755 / 15 0272 w - 80,000 lbs. u - -%- x .35:— - 30°02 1‘ 3° - 400000 in. 155. 1,42 . 400000 . 609 2/5 x' 2000 x .552: .755 53 - 509 x 1.5 d .. Gag" Say 10 inches, b - 5/8 x lO - 6 inches, D - 11 1/2 inches, A3 . 6 1:10 x.e0278 . 1.63 v ____ 400000. _ 554 5 x .755 :10 8 - one inch reinforcing bars are used. Web reinforcing is needed, therefore 3/8 inch steel is used, placed vertically every two inches throughout the length of the‘ beam. The beam was designed to be broken this necessitating the large value for fa. macs MADE 0N BRIQUETTES. Petoskey - New Aetna Neat Ottawa Sand Heat Ottawa sand. 280 360 180 195 185 315 185 235 180 190 230 180 230 325 235 250 350 220 305 165 240 345 160 255 220 260 330 310 350 285 325 180 295 296 200 250 .232 .522 .502 . .332 278 lbs. 278 lbs. per 235 lbs. 227 lbs. per per sq. in. sq. in. average ’ per sq. in.sq. in. average average average. TESTS.MADE 0N CYLINDERS BEFORE FIRING. smears! new mam 59.5000 45,000 g 4.2, 500 52, 400 45.500 41.800 45.900 44.500 43,175 average - 1542 lbs per sq. in. £4,125 average - 1676 lbs. per SQe ine TESTS MADE cn REINFORCED CONCRETE BEAMS. The beams that were not subjected to the fire test were loaded to failure in the Strength of Materials Labratory. They were found to fail by shear as our computations show. In both cases it was the concrete which failed and not the steel. The beams which were subjected to the fire test failed under very little load and when allowed to stand in the air for a few days all ambled up and fell to pieces. RESULTS of TESTS. Beam made with Petoskey cement. Failed at 2 6 570 lbs. Beam made with New Astana cement. Failed at 25980 lbs. The fire tests were conducted in the heat treatment laboratory of the Michigan Agricultural College. The furnace used was a large anealing furnace and an accurate means of recording the temperature by means of an electric thermo couple was used. The eight cylinders and fourty briquettes were placed in the same furnace at the same thus. It took 45 minutes to raise the temperature of the furnace from.60' F, to 1600' P, at which temperature it was kept for two hours. The specimens were then allowed to cool slowly and after a period of 24 hours were removed from the furnace. After a period of twenty minutes from the starting of the furnace cracks began to develop in the briquettes. Upon examination after the fire the briquettes were found to be very badly cracked. Some of the cracks extending through the specimen. These briquettes could stand no load and would crumble in the hands. They were also warped out cf‘shape. The cylinders were cracked but not as hadly as the briquettes.- When they were lifted by the hands the edges broke off'and crumbled, and when tested by the machine would not stand 1 1/2 lbs. per square inch. In some of the cylinders the concrete appeared to be fused on the surface. This was evident by the masses of material which had started to run down the sides. The reinforced beams were heated in the same furnace at a temperature of 1600°F, for two hours. The beams appeared to be as badly cracked as- the cylinders and upon being renoved from the furnace the edges crumbled not being able to support its own weight. f4 .' J A Concrete cylinder five days after jiring. \ Concrete beam as it looked five days after the fire, note the reinforcing iron. UR HUM i Q. we“ SPFLL/NG SKETCH 5HOW/N6 REINFORC/NG- STEEL dna’ FiflcTuRE CONCLUSIONS. Our very limited number of tests lead us to draw.the following conclusions. Concrete is far from being fire proof although to a large extent it is fire resisting. As far as we were able to determine there is no difference in the effect of fire on cement made from.Marl and that made from.limestcne. In all of our tests it‘uas the cement mortar rather than the aggregate that caused the concrete to fail. This is due to the hydration of the cement Which occurs at about 400°C, dissociating the cement into quick lime and water. The water is then evaporated and leaves the cement to crack and ultimately crumble. On two of the fired cylinders there was evidence of a slight fusion of the aggregate but in all cases the aggregate proved to be stronger than the mortar. The concrete seemed to be sufficiently strong enough during the firing but it is the after effect of the fire that causes the most damage. The importance of this "after effect of fire" cannot be over estimated. Even although it .might be possible to make a concrete which will stand its full load at the time of the conflagration, yet when the question.has to be faced as to whether the building is afterwards safe. and if not. how much of it should‘bc demolished and rebuilt, the answer would appear to be not very encouraging. In view of the very large number of buildings already created in reinforced concrete, the problem cannot safely be left at this point, but in the opinion of the writers the next work should be done on full sized specimens. The form of answer for practical conditions of work depends upon one or two further factors Which cannot satisfactorily be reproduced under laboratory conditions of the ordinary type. as far as the present results are concerned, it is submitted that reinforced concrete as at present carried out in practice is anything but fire proof and the temperature of primary importance is a comparatively low one. probably about 400°C, this being the approximate temperature of the dissociation of calcium hydrate. But as important a point as the resistance during fire is that of the after effects. It is suggested that extremely careful and very'skillful examination of reinforced concrete structures after a fire, is required, and it is highly probably that the original factor of safety can never be replaced in the building structure except by complete reconstruction. .10 qfl‘afll...‘ nan-(IUI . ......pr.§»a( ‘ 1.; a. , eL . . .. . . : a. .u a. £31.. n. .A\H..nuw h . ....AI..n.-HN.|I. - ...mu.1.\V;.JU|~r.q 1.3.3.... .1... Anew... .n e... .1... . c. . V. a t . u .. .x...r.|.,... . yr, . . ”LL inch u-,\K.~. Run-nuke 1.)...