Research Article The Prediction of Concrete Temperature ...
6
Hindawi Publishing Corporation Journal of Engineering Volume 2013, Article ID 946829, 5 pages http://dx.doi.org/10.1155/2013/946829 Research Article The Prediction of Concrete Temperature during Curing Using Regression and Artificial Neural Network Zahra Najafi 1 and Kaveh Ahangari 2 1 Department of Geology, Engineering Faculty, Science and Research Branch, Islamic Azad University, Poonak Square, Tehran, Iran 2 Department of Mining Engineering, Engineering Faculty, Science and Research Branch, Islamic Azad University, Toward Hesarak, End of Ashrafi Esfahani, Poonak Square, Tehran 1477893855, Iran Correspondence should be addressed to Kaveh Ahangari; [email protected] Received 5 December 2012; Accepted 12 February 2013 Academic Editor: ˙ Ilker B. Topc ¸u Copyright © 2013 Z. Najafi and K. Ahangari. is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Cement hydration plays a vital role in the temperature development of early-age concrete due to the heat generation. Concrete temperature affects the workability, and its measurement is an important element in any quality control program. In this regard, a method, which estimates the concrete temperature during curing, is very valuable. In this paper, multivariable regression and neural network methods were used for estimating concrete temperature. In order to achieve this purpose, ten laboratory cylindrical specimens were prepared under controlled situation, and concrete temperature was measured by thermistors existent in vibrating wire strain gauges. Input data variables consist of time (hour), environment temperature, water to cement ratio, aggregate content, height, and specimen diameter. Concrete temperature has been measured in ten different concrete specimens. Nonlinear regression achieved the determined coefficient (R 2 ) of 0.873. By using the same input set, the artificial neural network predicted concrete temperature with higher R 2 of 0.999. e results show that artificial neural network method significantly can be used to predict concrete temperature when regression results do not have appropriate accuracy. 1. Introduction Temperature prediction in fresh concrete is of great interest for designers and contractors because cement hydration is an exothermic process and the heat generation may lead to very early onset of thermal cracks in absence of any load [1]. erefore, utilizing a method that estimates temperature during curing is very beneficial. Cement hydration produces a rise in concrete internal temperature. Temperature rise varies by many parameters including cement composition, fineness and content, aggre- gate content and CTE (coefficient of thermal expansion), section geometry, placement, and ambient temperatures [2]. Aſter reaching the maximum temperature, the temperature of concrete decreases [3]. Pours with a large volume to surface area ratio are more susceptible to thermal cracking. Cements used for mass concrete should have a low C 3 S and C 3 A content to reduce excessive heat during hydration. Cement with a lower fineness with slow hydration reduces temperature rise. Mass concrete mixtures should contain as low of a cement content as possible to achieve the desired strength. is lowers the heat of hydration and subsequent temperature rise. A higher coarse aggregate content (70–85%) can be used to lower the cement content, reducing temperature rise. e CTE (coefficient of thermal expansion) of the coarse aggregate has the main influence on the CTE of the concrete. Lower CTE aggregates tend to have a higher thermal conductivity; thus, heat is released fast from the core. Lower ambient temperatures produce less temperature rise. Lower volume to surface ratio produces less temperature rise. / has a large effect on temperature rise. e lower / is, the less temperature rises [2]. Measuring concrete temperature during curing requires instrument and high costs. e used concrete tempera- ture prediction methods commonly consist of e Portland Cement Association (PCA) method, graphical method of ACI 207.2R, Schmidt’s method [4], and ConcreteWorks soſt- ware package [5].
Transcript of Research Article The Prediction of Concrete Temperature ...
Hindawi Publishing CorporationJournal of EngineeringVolume 2013 Article ID 946829 5 pageshttpdxdoiorg1011552013946829
Research ArticleThe Prediction of Concrete Temperature during CuringUsing Regression and Artificial Neural Network
Zahra Najafi1 and Kaveh Ahangari2
1 Department of Geology Engineering Faculty Science and Research Branch Islamic Azad University Poonak Square Tehran Iran2Department of Mining Engineering Engineering Faculty Science and Research Branch Islamic Azad UniversityToward Hesarak End of Ashrafi Esfahani Poonak Square Tehran 1477893855 Iran
Correspondence should be addressed to Kaveh Ahangari kavehahangarigmailcom
Received 5 December 2012 Accepted 12 February 2013
Academic Editor Ilker B Topcu
Copyright copy 2013 Z Najafi and K Ahangari This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
Cement hydration plays a vital role in the temperature development of early-age concrete due to the heat generation Concretetemperature affects the workability and its measurement is an important element in any quality control program In this regarda method which estimates the concrete temperature during curing is very valuable In this paper multivariable regression andneural networkmethods were used for estimating concrete temperature In order to achieve this purpose ten laboratory cylindricalspecimens were prepared under controlled situation and concrete temperature was measured by thermistors existent in vibratingwire strain gauges Input data variables consist of time (hour) environment temperature water to cement ratio aggregate contentheight and specimen diameter Concrete temperature has beenmeasured in ten different concrete specimens Nonlinear regressionachieved the determined coefficient (R2) of 0873 By using the same input set the artificial neural network predicted concretetemperature with higher R2 of 0999 The results show that artificial neural network method significantly can be used to predictconcrete temperature when regression results do not have appropriate accuracy
1 Introduction
Temperature prediction in fresh concrete is of great interestfor designers and contractors because cement hydration isan exothermic process and the heat generation may lead tovery early onset of thermal cracks in absence of any load[1] Therefore utilizing a method that estimates temperatureduring curing is very beneficial
Cement hydration produces a rise in concrete internaltemperature Temperature rise varies by many parametersincluding cement composition fineness and content aggre-gate content and CTE (coefficient of thermal expansion)section geometry placement and ambient temperatures [2]After reaching the maximum temperature the temperatureof concrete decreases [3]
Pours with a large volume to surface area ratio aremore susceptible to thermal cracking Cements used formass concrete should have a low C
3S and C
3A content to
reduce excessive heat during hydration Cement with a lowerfineness with slow hydration reduces temperature rise Mass
concrete mixtures should contain as low of a cement contentas possible to achieve the desired strength This lowers theheat of hydration and subsequent temperature rise A highercoarse aggregate content (70ndash85) can be used to lowerthe cement content reducing temperature rise The CTE(coefficient of thermal expansion) of the coarse aggregatehas the main influence on the CTE of the concrete LowerCTE aggregates tend to have a higher thermal conductivitythus heat is released fast from the core Lower ambienttemperatures produce less temperature rise Lower volumeto surface ratio produces less temperature rise 119882119862 has alarge effect on temperature rise The lower 119882119862 is the lesstemperature rises [2]
Measuring concrete temperature during curing requiresinstrument and high costs The used concrete tempera-ture prediction methods commonly consist of The PortlandCement Association (PCA) method graphical method ofACI 2072R Schmidtrsquos method [4] and ConcreteWorks soft-ware package [5]
2 Journal of Engineering
The PCA Method calculates 10∘F temperature rise forevery 100 lb of cement provides no information on timeof maximum temperature does not allow the quantifica-tion of temperature differences and assumes that the leastdimension of the concrete member is at least 18m (6 ft)Graphical method of ACI 2072R uses charts and equationsbased on empirical data and assumptions are for boundaryconditionsGenerally thismethodunderestimatesmaximumtemperature and is poor predictor of time to achieve max-imum temperature Schmidtrsquos method is little guidance forboundary conditions and difficult to model Moreover it canbe complicated and should be performed by an experiencedengineer [5] In addition to defects of three above-mentionedmethods they do not predict continuous concrete temper-ature ConcreteWorks Software package used for predictingcontinuous concrete temperature needs to measure amountsof concrete air content slump specified final compressivestrength (1198911015840
119888) coefficient of concrete thermal expansion
and thermal properties This type of measuring spends toomuch time and cost Thus using quick and easy methodfor prediction of continuous concrete temperature whichmeasures input parameters in an easy and inexpensive waycould be very useful
The aim of this study is predicting the temperature duringconcrete curing by use of time (ℎ) environment temperaturewater to cement ratio aggregate content diameter andspecimen height as variables The required data are a resultof laboratory experiment Multivariate regression (SPSS soft-ware) and artificial neural network (MATLAB) have beenused for prediction
2 Experimental Procedures
In order to predict temperature during concrete curing itis required to measure the temperature continuously usingthe thermistors which are located inside the concrete sam-ples The necessary data is obtained from ten experimentscarried out on different cylindrical concrete specimens in theInstitute of Geotechnical Engineering andMine Surveying ofTechnical University of Clausthal Germany Different typesof vibrating wire strain gauges were installed in each concretespecimenThe vibrating wire strain gauges are equipped withthermistors and the concrete temperature was measured byit During different stages of concreting in the concrete wasappropriately compacted by a manual vibrator
Themeasuring began right after specimen concreting andduring the curing process Temperature was recorded until 30hours after concreting which the temperature changes wererather stopped
In order to predict the temperaturemore accurately mea-sured concrete temperature in specimens with similar straingauges possibility was utilized as concrete temperatures
The type of cement used in this study and produced byGerman Deuna Co is a Portland cement (CEM I 425R)
The characteristics of specimens are presented in Table 1The used aggregates in all specimens are coarse and silicatype Specimen no 9 was put in cold weather (minus2∘C until+184∘C) after concreting and during the curing process For
Table 1 Characteristics of experimental specimens for set 119868
Specimenno 119882119862 () Diameter
(mm)Height(mm)
Aggregate(Kg)
1 50 460 480 914582 50 300 480 914583 50 200 480 914584 67 300 480 445 50 300 250 877726 50 200 250 877727 67 460 480 42488 61 460 480 433279 65 460 480 8810 50 460 480 87772
0
5
10
15
20
25
30
35
40
0 5 10 15 20 25 30 35Time (h)
Concrete temperature (specimen 1) Concrete temperature (specimen 2)
Concrete temperature (specimen 3) Concrete temperature (specimen 4)Concrete temperature (specimen 5) Concrete temperature (specimen 6)Concrete temperature (specimen 7) Concrete temperature (specimen 8)
Concrete temperature (specimen 9) Concrete temperature (specimen 10)
Con
cret
e tem
pera
ture
(∘C)
Figure 1Measured temperature changes during curing for concretespecimens
specimens with water to cement ratio of 50 and specimenno 9 (concreting in cold weather) 30mL plasticizer was usedfor each kilogram of cement
The measured temperature changes during curing forspecimens are presented in Figure 1
3 Data Analysis and the Results
31 Multivariable Regression In this study both linearand nonlinear regressions were used to develop equationsbetween concrete temperature and input variables Thestepwise variable selection procedure was applied to prepareequationsThe statistical parameters of the input variables areshown in Table 2
By using the least squaremathematical method the inter-correlations of time (ℎ) environment temperature water tocement ratio (119882119862) aggregate content specimen height anddiameter with concrete temperature were calculated at 04860704 0181 minus0617 0032 and 0228 respectively The resultsshow that with the increase of environment temperature and
Journal of Engineering 3
Table 2 The range of variables
Variable () Minimum Maximum Mean Standard deviation Number of dataTime (hour) 000 30433 1012 9049 2340Environment temperature (∘C) minus2 2248 15496 6142 2340Watercement ratio () 50 67 5862 736 2340Aggregate weight (kg) 4248 91458 6717 2314 2340Diameter (mm) 200 460 39668 9657 2340Height (mm) 250 480 45258 7455 2340
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0420minus2minus4
Std Dev = 0999Mean = 728119864minus15
119873 = 2340
Figure 2 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (1)
time lapse concrete temperature rises and with the increaseof aggregate content concrete temperature decreases Theeffect of other parameters on concrete temperature is notsignificant
The linear equation between input variables and concretetemperature is as follows
119879 = 12425 + 032119905 + 0694119879out minus 0064 (119882
119862)
minus 0107119892 + 0028119889 minus 0018ℎ 1198772= 0814
(1)
In addition the nonlinear equation between parameters is asfollows
119879 = minus 2195 + 0946 119905 + 1731119879out minus 0356 (119882
119862)
+ 1705119892 + 002119889 minus 0007ℎ minus 00261199052minus 0203119879
2
out
minus 00141198922+ 0007119879
3
out 1198772= 0873
(2)
in which 119905 119879out 119892 (119882119862) 119889 and ℎ are time (ℎ) environmenttemperature (∘C) aggregate amount (Kg) water to cementratio concrete specimen diameter (mm) and height (mm)respectively
The distribution of the difference between concrete tem-perature predicted from (1) and (2) and actual determined
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0151050minus5minus10
Std Dev = 2809Mean = 792119864minus11
119873 = 2340
Figure 3 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (2)
amounts are shown in Figures 2 and 3 The results indicatethat (2) can have a significant estimation of concrete temper-ature during curing for concrete made from CEM I 425R
32 Artificial Neural Network Procedure Among the existingnumerous neural networks (NNs) paradigms feed-forwardartificial neural networks (FANNs) are the most popular dueto their flexibility in structure good presentational capabili-ties and large number of available training algorithms [6ndash8]
The basic structure of a multilayer feed-forward networkmodel can be made of one input layer one or more hiddenlayers and one output layer [9]
Neural network training can be made more efficient byspecific preprocessing In this paper all the input and outputparameters were preprocessed by normalizing inputs andtargets therefore in the preprocessing stage their mean andstandard deviation are 0 and 1 respectively Consider thefollowing
119873119901=
(119860119901minusmean119860
119901119904)
std 119860119901
(3)
in which 119860119901is an actual parameter mean 119860
119901119904is actual
parameters mean std 119860119901
is actual parameters standarddeviation and119873
119901is a normalized parameter [10]
4 Journal of Engineering
Table 3 Details of ANN model
Input sets Trainingset size
Testing setsize
Validationset size 119868 119869 119870
119905 119879out (119882119862)119892 119889 ℎ 1404 468 468 5 6 6
119868 number of input nodes 119869 number of nodes in the first hidden layer and119870 number of nodes in the second hidden layer
Table 4 Statistical analysis of predicted temperature and thegeneralized performance of ANN-based model
Validation stage Training stageCorrelation coefficient 09996 09993
In this part of the study ANN model is presented topredict concrete temperature during curing Multilayer feed-forward network model has been trained with BP (backpropagation) training algorithm
Different neural networks were designed and the bestparameters value was obtained by trial and error Howeverthe main aim is to acquire a neural network with the smallestdimensions and the least errorsThemost appropriate resultshave been obtained from chosen network model in whichhyperbolic tangent sigmoid and linear functions were usedas an activation function for the hidden and output layer neu-rons According to (1) the selected variables were determinedas the best variables for predicting concrete temperatureTherefore those variables which were used as input to ANNfor the improvement of concrete temperature prediction arelisted in Table 3
The data in the model were separated into three trainingvalidation and test sets in which the test set was used aftertraining The validation and training process were stoppedafter 245 epochs for model
The performance function is the mean square error(MSE) the average-squared error between the network pre-dicted outputs and the target outputs which are equal to000044 for training The correlation coefficients for thevalidation and training stages are presented in Table 4
Figures 4 5 and 6 show a graphical comparison of thedetermined experimental temperature and those predicted byartificial neural network in the validation training and testprocess for model
The distribution of the difference between predictedtemperature by ANN and actual values in the test process ispresented in Figure 7
It was observed that concrete temperature predictionusing ANN procedure could be more acceptable and satis-factory than others
4 Conclusions
(i) Concrete temperature during curing was studied byexperimenting on ten cylindrical concrete specimensunder controlled situations in which the concretetemperature was measured by strain gauges (whichare equipped with thermistor) Data recording began
0 50 100 150 200 250 300 350 400 450 50005
101520253035
Data number
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09996
Figure 4 Graphical comparison of temperature with those pre-dicted by ANN in the validation process
0 500 1000 1500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09993
Figure 5 Graphical comparison of temperature with those pre-dicted by ANN in the training process
0 50 100 150 200 250 300 350 400 450 500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 0999
Figure 6 Graphical comparison of temperature with those pre-dicted by ANN in the test process
right after specimen concreting and continued until30 hours after
(ii) The model used for concrete specimens which wereprepared with Portland cement (CEM I 425R) wasutilized for estimating concrete temperature usingstepwise regression and artificial neural networkmethods
Journal of Engineering 5
Difference between actual and ANN predicted temperature
Freq
uenc
y
200
150
100
50
0210minus1minus2minus3
Mean = 0Std Dev = 021119873 = 468
Figure 7 Distribution of the difference between predicted temper-ature by ANN and actual values in the test process
(iii) The inter correlation between input variables andconcrete temperature showed that higher environ-ment temperature and time in concrete can resultin higher concrete temperature and higher aggregatecontent in concrete results in lower-concrete temper-ature No other parameters were significant
(iv) The linear and nonlinear equations can estimate theconcrete temperature with correlation coefficients(1198772) of 0814 and 0873 respectively
(v) The artificial neural network procedure can predictthe concrete temperature with correlation coefficientof 0999 Therefore the obtained results are muchbetter than multivariate regression
(vi) The results show that artificial neural network isa reliable method to predict concrete temperatureduring curing
Conflict of Interests
Hereby the authors disclosure that this paper was just acontribution to the advancement of science and they justused SPSS and MATLAB softwares as the mathematicalmethods for prediction The authors do not have a directfinancial relation with the commercial identity mentioned inthe paper (SPSS and MATLAB softwares)
References
[1] P F Siew T Puapansawat and YHWu ldquoTemperature and heatstress in a concrete column at early agesrdquoANZIAM Journal vol44 no E pp C705ndashC722 2003
[2] R Moser ldquoMass Concrete CEE8813AmdashMaterial scienceof concrete 2 Lecture Overviewrdquo httppeoplecegatechedusimkk92massconcretepdf
[3] H Weigler and S Karl Junger Beton Beanspruchung-Festigkeit-Verformung vol 40 Betonwerk Fertigteil-Technik 1974
[4] K A Riding J L Poole A K Schindler M C G Juenger andK J Folliard ldquoEvaluation of temperature prediction methodsformass concretemembersrdquoACIMaterials Journal vol 103 no5 pp 357ndash365 2006
[5] ConcreteWorks IHEEP Conference San Antonio Tex USA2009
[6] C T LeondesNeural Network Systems Techniques and Applica-tions Algorithms and Architectures Academic Press New YorkNY USA 1998
[7] R P Lippmann ldquoAn introduction to computing with neuralnetsrdquo IEEE ASSP Magazine vol 4 no 2 pp 4ndash22 1987
[8] D Sarkar ldquoMethods to speed up error back-propagation learn-ing algorithmrdquo ACMComputing Surveys vol 27 no 4 pp 519ndash542 1995
[9] F Ozcan CDAtis OKarahan EUncuoglu andHTanyildizildquoComparison of artificial neural network and fuzzy logic mod-els for prediction of long-term compressive strength of silicafume concreterdquo Advances in Engineering Software vol 40 no9 pp 856ndash863 2009
[10] H Demuth and M Beale Neural Network Toolbox for Use withMATLAB Handbook 2002
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
2 Journal of Engineering
The PCA Method calculates 10∘F temperature rise forevery 100 lb of cement provides no information on timeof maximum temperature does not allow the quantifica-tion of temperature differences and assumes that the leastdimension of the concrete member is at least 18m (6 ft)Graphical method of ACI 2072R uses charts and equationsbased on empirical data and assumptions are for boundaryconditionsGenerally thismethodunderestimatesmaximumtemperature and is poor predictor of time to achieve max-imum temperature Schmidtrsquos method is little guidance forboundary conditions and difficult to model Moreover it canbe complicated and should be performed by an experiencedengineer [5] In addition to defects of three above-mentionedmethods they do not predict continuous concrete temper-ature ConcreteWorks Software package used for predictingcontinuous concrete temperature needs to measure amountsof concrete air content slump specified final compressivestrength (1198911015840
119888) coefficient of concrete thermal expansion
and thermal properties This type of measuring spends toomuch time and cost Thus using quick and easy methodfor prediction of continuous concrete temperature whichmeasures input parameters in an easy and inexpensive waycould be very useful
The aim of this study is predicting the temperature duringconcrete curing by use of time (ℎ) environment temperaturewater to cement ratio aggregate content diameter andspecimen height as variables The required data are a resultof laboratory experiment Multivariate regression (SPSS soft-ware) and artificial neural network (MATLAB) have beenused for prediction
2 Experimental Procedures
In order to predict temperature during concrete curing itis required to measure the temperature continuously usingthe thermistors which are located inside the concrete sam-ples The necessary data is obtained from ten experimentscarried out on different cylindrical concrete specimens in theInstitute of Geotechnical Engineering andMine Surveying ofTechnical University of Clausthal Germany Different typesof vibrating wire strain gauges were installed in each concretespecimenThe vibrating wire strain gauges are equipped withthermistors and the concrete temperature was measured byit During different stages of concreting in the concrete wasappropriately compacted by a manual vibrator
Themeasuring began right after specimen concreting andduring the curing process Temperature was recorded until 30hours after concreting which the temperature changes wererather stopped
In order to predict the temperaturemore accurately mea-sured concrete temperature in specimens with similar straingauges possibility was utilized as concrete temperatures
The type of cement used in this study and produced byGerman Deuna Co is a Portland cement (CEM I 425R)
The characteristics of specimens are presented in Table 1The used aggregates in all specimens are coarse and silicatype Specimen no 9 was put in cold weather (minus2∘C until+184∘C) after concreting and during the curing process For
Table 1 Characteristics of experimental specimens for set 119868
Specimenno 119882119862 () Diameter
(mm)Height(mm)
Aggregate(Kg)
1 50 460 480 914582 50 300 480 914583 50 200 480 914584 67 300 480 445 50 300 250 877726 50 200 250 877727 67 460 480 42488 61 460 480 433279 65 460 480 8810 50 460 480 87772
0
5
10
15
20
25
30
35
40
0 5 10 15 20 25 30 35Time (h)
Concrete temperature (specimen 1) Concrete temperature (specimen 2)
Concrete temperature (specimen 3) Concrete temperature (specimen 4)Concrete temperature (specimen 5) Concrete temperature (specimen 6)Concrete temperature (specimen 7) Concrete temperature (specimen 8)
Concrete temperature (specimen 9) Concrete temperature (specimen 10)
Con
cret
e tem
pera
ture
(∘C)
Figure 1Measured temperature changes during curing for concretespecimens
specimens with water to cement ratio of 50 and specimenno 9 (concreting in cold weather) 30mL plasticizer was usedfor each kilogram of cement
The measured temperature changes during curing forspecimens are presented in Figure 1
3 Data Analysis and the Results
31 Multivariable Regression In this study both linearand nonlinear regressions were used to develop equationsbetween concrete temperature and input variables Thestepwise variable selection procedure was applied to prepareequationsThe statistical parameters of the input variables areshown in Table 2
By using the least squaremathematical method the inter-correlations of time (ℎ) environment temperature water tocement ratio (119882119862) aggregate content specimen height anddiameter with concrete temperature were calculated at 04860704 0181 minus0617 0032 and 0228 respectively The resultsshow that with the increase of environment temperature and
Journal of Engineering 3
Table 2 The range of variables
Variable () Minimum Maximum Mean Standard deviation Number of dataTime (hour) 000 30433 1012 9049 2340Environment temperature (∘C) minus2 2248 15496 6142 2340Watercement ratio () 50 67 5862 736 2340Aggregate weight (kg) 4248 91458 6717 2314 2340Diameter (mm) 200 460 39668 9657 2340Height (mm) 250 480 45258 7455 2340
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0420minus2minus4
Std Dev = 0999Mean = 728119864minus15
119873 = 2340
Figure 2 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (1)
time lapse concrete temperature rises and with the increaseof aggregate content concrete temperature decreases Theeffect of other parameters on concrete temperature is notsignificant
The linear equation between input variables and concretetemperature is as follows
119879 = 12425 + 032119905 + 0694119879out minus 0064 (119882
119862)
minus 0107119892 + 0028119889 minus 0018ℎ 1198772= 0814
(1)
In addition the nonlinear equation between parameters is asfollows
119879 = minus 2195 + 0946 119905 + 1731119879out minus 0356 (119882
119862)
+ 1705119892 + 002119889 minus 0007ℎ minus 00261199052minus 0203119879
2
out
minus 00141198922+ 0007119879
3
out 1198772= 0873
(2)
in which 119905 119879out 119892 (119882119862) 119889 and ℎ are time (ℎ) environmenttemperature (∘C) aggregate amount (Kg) water to cementratio concrete specimen diameter (mm) and height (mm)respectively
The distribution of the difference between concrete tem-perature predicted from (1) and (2) and actual determined
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0151050minus5minus10
Std Dev = 2809Mean = 792119864minus11
119873 = 2340
Figure 3 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (2)
amounts are shown in Figures 2 and 3 The results indicatethat (2) can have a significant estimation of concrete temper-ature during curing for concrete made from CEM I 425R
32 Artificial Neural Network Procedure Among the existingnumerous neural networks (NNs) paradigms feed-forwardartificial neural networks (FANNs) are the most popular dueto their flexibility in structure good presentational capabili-ties and large number of available training algorithms [6ndash8]
The basic structure of a multilayer feed-forward networkmodel can be made of one input layer one or more hiddenlayers and one output layer [9]
Neural network training can be made more efficient byspecific preprocessing In this paper all the input and outputparameters were preprocessed by normalizing inputs andtargets therefore in the preprocessing stage their mean andstandard deviation are 0 and 1 respectively Consider thefollowing
119873119901=
(119860119901minusmean119860
119901119904)
std 119860119901
(3)
in which 119860119901is an actual parameter mean 119860
119901119904is actual
parameters mean std 119860119901
is actual parameters standarddeviation and119873
119901is a normalized parameter [10]
4 Journal of Engineering
Table 3 Details of ANN model
Input sets Trainingset size
Testing setsize
Validationset size 119868 119869 119870
119905 119879out (119882119862)119892 119889 ℎ 1404 468 468 5 6 6
119868 number of input nodes 119869 number of nodes in the first hidden layer and119870 number of nodes in the second hidden layer
Table 4 Statistical analysis of predicted temperature and thegeneralized performance of ANN-based model
Validation stage Training stageCorrelation coefficient 09996 09993
In this part of the study ANN model is presented topredict concrete temperature during curing Multilayer feed-forward network model has been trained with BP (backpropagation) training algorithm
Different neural networks were designed and the bestparameters value was obtained by trial and error Howeverthe main aim is to acquire a neural network with the smallestdimensions and the least errorsThemost appropriate resultshave been obtained from chosen network model in whichhyperbolic tangent sigmoid and linear functions were usedas an activation function for the hidden and output layer neu-rons According to (1) the selected variables were determinedas the best variables for predicting concrete temperatureTherefore those variables which were used as input to ANNfor the improvement of concrete temperature prediction arelisted in Table 3
The data in the model were separated into three trainingvalidation and test sets in which the test set was used aftertraining The validation and training process were stoppedafter 245 epochs for model
The performance function is the mean square error(MSE) the average-squared error between the network pre-dicted outputs and the target outputs which are equal to000044 for training The correlation coefficients for thevalidation and training stages are presented in Table 4
Figures 4 5 and 6 show a graphical comparison of thedetermined experimental temperature and those predicted byartificial neural network in the validation training and testprocess for model
The distribution of the difference between predictedtemperature by ANN and actual values in the test process ispresented in Figure 7
It was observed that concrete temperature predictionusing ANN procedure could be more acceptable and satis-factory than others
4 Conclusions
(i) Concrete temperature during curing was studied byexperimenting on ten cylindrical concrete specimensunder controlled situations in which the concretetemperature was measured by strain gauges (whichare equipped with thermistor) Data recording began
0 50 100 150 200 250 300 350 400 450 50005
101520253035
Data number
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09996
Figure 4 Graphical comparison of temperature with those pre-dicted by ANN in the validation process
0 500 1000 1500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09993
Figure 5 Graphical comparison of temperature with those pre-dicted by ANN in the training process
0 50 100 150 200 250 300 350 400 450 500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 0999
Figure 6 Graphical comparison of temperature with those pre-dicted by ANN in the test process
right after specimen concreting and continued until30 hours after
(ii) The model used for concrete specimens which wereprepared with Portland cement (CEM I 425R) wasutilized for estimating concrete temperature usingstepwise regression and artificial neural networkmethods
Journal of Engineering 5
Difference between actual and ANN predicted temperature
Freq
uenc
y
200
150
100
50
0210minus1minus2minus3
Mean = 0Std Dev = 021119873 = 468
Figure 7 Distribution of the difference between predicted temper-ature by ANN and actual values in the test process
(iii) The inter correlation between input variables andconcrete temperature showed that higher environ-ment temperature and time in concrete can resultin higher concrete temperature and higher aggregatecontent in concrete results in lower-concrete temper-ature No other parameters were significant
(iv) The linear and nonlinear equations can estimate theconcrete temperature with correlation coefficients(1198772) of 0814 and 0873 respectively
(v) The artificial neural network procedure can predictthe concrete temperature with correlation coefficientof 0999 Therefore the obtained results are muchbetter than multivariate regression
(vi) The results show that artificial neural network isa reliable method to predict concrete temperatureduring curing
Conflict of Interests
Hereby the authors disclosure that this paper was just acontribution to the advancement of science and they justused SPSS and MATLAB softwares as the mathematicalmethods for prediction The authors do not have a directfinancial relation with the commercial identity mentioned inthe paper (SPSS and MATLAB softwares)
References
[1] P F Siew T Puapansawat and YHWu ldquoTemperature and heatstress in a concrete column at early agesrdquoANZIAM Journal vol44 no E pp C705ndashC722 2003
[2] R Moser ldquoMass Concrete CEE8813AmdashMaterial scienceof concrete 2 Lecture Overviewrdquo httppeoplecegatechedusimkk92massconcretepdf
[3] H Weigler and S Karl Junger Beton Beanspruchung-Festigkeit-Verformung vol 40 Betonwerk Fertigteil-Technik 1974
[4] K A Riding J L Poole A K Schindler M C G Juenger andK J Folliard ldquoEvaluation of temperature prediction methodsformass concretemembersrdquoACIMaterials Journal vol 103 no5 pp 357ndash365 2006
[5] ConcreteWorks IHEEP Conference San Antonio Tex USA2009
[6] C T LeondesNeural Network Systems Techniques and Applica-tions Algorithms and Architectures Academic Press New YorkNY USA 1998
[7] R P Lippmann ldquoAn introduction to computing with neuralnetsrdquo IEEE ASSP Magazine vol 4 no 2 pp 4ndash22 1987
[8] D Sarkar ldquoMethods to speed up error back-propagation learn-ing algorithmrdquo ACMComputing Surveys vol 27 no 4 pp 519ndash542 1995
[9] F Ozcan CDAtis OKarahan EUncuoglu andHTanyildizildquoComparison of artificial neural network and fuzzy logic mod-els for prediction of long-term compressive strength of silicafume concreterdquo Advances in Engineering Software vol 40 no9 pp 856ndash863 2009
[10] H Demuth and M Beale Neural Network Toolbox for Use withMATLAB Handbook 2002
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 3
Table 2 The range of variables
Variable () Minimum Maximum Mean Standard deviation Number of dataTime (hour) 000 30433 1012 9049 2340Environment temperature (∘C) minus2 2248 15496 6142 2340Watercement ratio () 50 67 5862 736 2340Aggregate weight (kg) 4248 91458 6717 2314 2340Diameter (mm) 200 460 39668 9657 2340Height (mm) 250 480 45258 7455 2340
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0420minus2minus4
Std Dev = 0999Mean = 728119864minus15
119873 = 2340
Figure 2 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (1)
time lapse concrete temperature rises and with the increaseof aggregate content concrete temperature decreases Theeffect of other parameters on concrete temperature is notsignificant
The linear equation between input variables and concretetemperature is as follows
119879 = 12425 + 032119905 + 0694119879out minus 0064 (119882
119862)
minus 0107119892 + 0028119889 minus 0018ℎ 1198772= 0814
(1)
In addition the nonlinear equation between parameters is asfollows
119879 = minus 2195 + 0946 119905 + 1731119879out minus 0356 (119882
119862)
+ 1705119892 + 002119889 minus 0007ℎ minus 00261199052minus 0203119879
2
out
minus 00141198922+ 0007119879
3
out 1198772= 0873
(2)
in which 119905 119879out 119892 (119882119862) 119889 and ℎ are time (ℎ) environmenttemperature (∘C) aggregate amount (Kg) water to cementratio concrete specimen diameter (mm) and height (mm)respectively
The distribution of the difference between concrete tem-perature predicted from (1) and (2) and actual determined
Difference between actual and predicted temperature
Freq
uenc
y
250
200
150
100
50
0151050minus5minus10
Std Dev = 2809Mean = 792119864minus11
119873 = 2340
Figure 3 Distribution of the difference between actual temperaturevalues and predicted temperature values obtained frommultivariateregression (2)
amounts are shown in Figures 2 and 3 The results indicatethat (2) can have a significant estimation of concrete temper-ature during curing for concrete made from CEM I 425R
32 Artificial Neural Network Procedure Among the existingnumerous neural networks (NNs) paradigms feed-forwardartificial neural networks (FANNs) are the most popular dueto their flexibility in structure good presentational capabili-ties and large number of available training algorithms [6ndash8]
The basic structure of a multilayer feed-forward networkmodel can be made of one input layer one or more hiddenlayers and one output layer [9]
Neural network training can be made more efficient byspecific preprocessing In this paper all the input and outputparameters were preprocessed by normalizing inputs andtargets therefore in the preprocessing stage their mean andstandard deviation are 0 and 1 respectively Consider thefollowing
119873119901=
(119860119901minusmean119860
119901119904)
std 119860119901
(3)
in which 119860119901is an actual parameter mean 119860
119901119904is actual
parameters mean std 119860119901
is actual parameters standarddeviation and119873
119901is a normalized parameter [10]
4 Journal of Engineering
Table 3 Details of ANN model
Input sets Trainingset size
Testing setsize
Validationset size 119868 119869 119870
119905 119879out (119882119862)119892 119889 ℎ 1404 468 468 5 6 6
119868 number of input nodes 119869 number of nodes in the first hidden layer and119870 number of nodes in the second hidden layer
Table 4 Statistical analysis of predicted temperature and thegeneralized performance of ANN-based model
Validation stage Training stageCorrelation coefficient 09996 09993
In this part of the study ANN model is presented topredict concrete temperature during curing Multilayer feed-forward network model has been trained with BP (backpropagation) training algorithm
Different neural networks were designed and the bestparameters value was obtained by trial and error Howeverthe main aim is to acquire a neural network with the smallestdimensions and the least errorsThemost appropriate resultshave been obtained from chosen network model in whichhyperbolic tangent sigmoid and linear functions were usedas an activation function for the hidden and output layer neu-rons According to (1) the selected variables were determinedas the best variables for predicting concrete temperatureTherefore those variables which were used as input to ANNfor the improvement of concrete temperature prediction arelisted in Table 3
The data in the model were separated into three trainingvalidation and test sets in which the test set was used aftertraining The validation and training process were stoppedafter 245 epochs for model
The performance function is the mean square error(MSE) the average-squared error between the network pre-dicted outputs and the target outputs which are equal to000044 for training The correlation coefficients for thevalidation and training stages are presented in Table 4
Figures 4 5 and 6 show a graphical comparison of thedetermined experimental temperature and those predicted byartificial neural network in the validation training and testprocess for model
The distribution of the difference between predictedtemperature by ANN and actual values in the test process ispresented in Figure 7
It was observed that concrete temperature predictionusing ANN procedure could be more acceptable and satis-factory than others
4 Conclusions
(i) Concrete temperature during curing was studied byexperimenting on ten cylindrical concrete specimensunder controlled situations in which the concretetemperature was measured by strain gauges (whichare equipped with thermistor) Data recording began
0 50 100 150 200 250 300 350 400 450 50005
101520253035
Data number
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09996
Figure 4 Graphical comparison of temperature with those pre-dicted by ANN in the validation process
0 500 1000 1500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09993
Figure 5 Graphical comparison of temperature with those pre-dicted by ANN in the training process
0 50 100 150 200 250 300 350 400 450 500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 0999
Figure 6 Graphical comparison of temperature with those pre-dicted by ANN in the test process
right after specimen concreting and continued until30 hours after
(ii) The model used for concrete specimens which wereprepared with Portland cement (CEM I 425R) wasutilized for estimating concrete temperature usingstepwise regression and artificial neural networkmethods
Journal of Engineering 5
Difference between actual and ANN predicted temperature
Freq
uenc
y
200
150
100
50
0210minus1minus2minus3
Mean = 0Std Dev = 021119873 = 468
Figure 7 Distribution of the difference between predicted temper-ature by ANN and actual values in the test process
(iii) The inter correlation between input variables andconcrete temperature showed that higher environ-ment temperature and time in concrete can resultin higher concrete temperature and higher aggregatecontent in concrete results in lower-concrete temper-ature No other parameters were significant
(iv) The linear and nonlinear equations can estimate theconcrete temperature with correlation coefficients(1198772) of 0814 and 0873 respectively
(v) The artificial neural network procedure can predictthe concrete temperature with correlation coefficientof 0999 Therefore the obtained results are muchbetter than multivariate regression
(vi) The results show that artificial neural network isa reliable method to predict concrete temperatureduring curing
Conflict of Interests
Hereby the authors disclosure that this paper was just acontribution to the advancement of science and they justused SPSS and MATLAB softwares as the mathematicalmethods for prediction The authors do not have a directfinancial relation with the commercial identity mentioned inthe paper (SPSS and MATLAB softwares)
References
[1] P F Siew T Puapansawat and YHWu ldquoTemperature and heatstress in a concrete column at early agesrdquoANZIAM Journal vol44 no E pp C705ndashC722 2003
[2] R Moser ldquoMass Concrete CEE8813AmdashMaterial scienceof concrete 2 Lecture Overviewrdquo httppeoplecegatechedusimkk92massconcretepdf
[3] H Weigler and S Karl Junger Beton Beanspruchung-Festigkeit-Verformung vol 40 Betonwerk Fertigteil-Technik 1974
[4] K A Riding J L Poole A K Schindler M C G Juenger andK J Folliard ldquoEvaluation of temperature prediction methodsformass concretemembersrdquoACIMaterials Journal vol 103 no5 pp 357ndash365 2006
[5] ConcreteWorks IHEEP Conference San Antonio Tex USA2009
[6] C T LeondesNeural Network Systems Techniques and Applica-tions Algorithms and Architectures Academic Press New YorkNY USA 1998
[7] R P Lippmann ldquoAn introduction to computing with neuralnetsrdquo IEEE ASSP Magazine vol 4 no 2 pp 4ndash22 1987
[8] D Sarkar ldquoMethods to speed up error back-propagation learn-ing algorithmrdquo ACMComputing Surveys vol 27 no 4 pp 519ndash542 1995
[9] F Ozcan CDAtis OKarahan EUncuoglu andHTanyildizildquoComparison of artificial neural network and fuzzy logic mod-els for prediction of long-term compressive strength of silicafume concreterdquo Advances in Engineering Software vol 40 no9 pp 856ndash863 2009
[10] H Demuth and M Beale Neural Network Toolbox for Use withMATLAB Handbook 2002
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
4 Journal of Engineering
Table 3 Details of ANN model
Input sets Trainingset size
Testing setsize
Validationset size 119868 119869 119870
119905 119879out (119882119862)119892 119889 ℎ 1404 468 468 5 6 6
119868 number of input nodes 119869 number of nodes in the first hidden layer and119870 number of nodes in the second hidden layer
Table 4 Statistical analysis of predicted temperature and thegeneralized performance of ANN-based model
Validation stage Training stageCorrelation coefficient 09996 09993
In this part of the study ANN model is presented topredict concrete temperature during curing Multilayer feed-forward network model has been trained with BP (backpropagation) training algorithm
Different neural networks were designed and the bestparameters value was obtained by trial and error Howeverthe main aim is to acquire a neural network with the smallestdimensions and the least errorsThemost appropriate resultshave been obtained from chosen network model in whichhyperbolic tangent sigmoid and linear functions were usedas an activation function for the hidden and output layer neu-rons According to (1) the selected variables were determinedas the best variables for predicting concrete temperatureTherefore those variables which were used as input to ANNfor the improvement of concrete temperature prediction arelisted in Table 3
The data in the model were separated into three trainingvalidation and test sets in which the test set was used aftertraining The validation and training process were stoppedafter 245 epochs for model
The performance function is the mean square error(MSE) the average-squared error between the network pre-dicted outputs and the target outputs which are equal to000044 for training The correlation coefficients for thevalidation and training stages are presented in Table 4
Figures 4 5 and 6 show a graphical comparison of thedetermined experimental temperature and those predicted byartificial neural network in the validation training and testprocess for model
The distribution of the difference between predictedtemperature by ANN and actual values in the test process ispresented in Figure 7
It was observed that concrete temperature predictionusing ANN procedure could be more acceptable and satis-factory than others
4 Conclusions
(i) Concrete temperature during curing was studied byexperimenting on ten cylindrical concrete specimensunder controlled situations in which the concretetemperature was measured by strain gauges (whichare equipped with thermistor) Data recording began
0 50 100 150 200 250 300 350 400 450 50005
101520253035
Data number
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09996
Figure 4 Graphical comparison of temperature with those pre-dicted by ANN in the validation process
0 500 1000 1500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 09993
Figure 5 Graphical comparison of temperature with those pre-dicted by ANN in the training process
0 50 100 150 200 250 300 350 400 450 500Data number
05
101520253035
Con
cret
e tem
pera
ture
ActualANN predicted
1198772= 0999
Figure 6 Graphical comparison of temperature with those pre-dicted by ANN in the test process
right after specimen concreting and continued until30 hours after
(ii) The model used for concrete specimens which wereprepared with Portland cement (CEM I 425R) wasutilized for estimating concrete temperature usingstepwise regression and artificial neural networkmethods
Journal of Engineering 5
Difference between actual and ANN predicted temperature
Freq
uenc
y
200
150
100
50
0210minus1minus2minus3
Mean = 0Std Dev = 021119873 = 468
Figure 7 Distribution of the difference between predicted temper-ature by ANN and actual values in the test process
(iii) The inter correlation between input variables andconcrete temperature showed that higher environ-ment temperature and time in concrete can resultin higher concrete temperature and higher aggregatecontent in concrete results in lower-concrete temper-ature No other parameters were significant
(iv) The linear and nonlinear equations can estimate theconcrete temperature with correlation coefficients(1198772) of 0814 and 0873 respectively
(v) The artificial neural network procedure can predictthe concrete temperature with correlation coefficientof 0999 Therefore the obtained results are muchbetter than multivariate regression
(vi) The results show that artificial neural network isa reliable method to predict concrete temperatureduring curing
Conflict of Interests
Hereby the authors disclosure that this paper was just acontribution to the advancement of science and they justused SPSS and MATLAB softwares as the mathematicalmethods for prediction The authors do not have a directfinancial relation with the commercial identity mentioned inthe paper (SPSS and MATLAB softwares)
References
[1] P F Siew T Puapansawat and YHWu ldquoTemperature and heatstress in a concrete column at early agesrdquoANZIAM Journal vol44 no E pp C705ndashC722 2003
[2] R Moser ldquoMass Concrete CEE8813AmdashMaterial scienceof concrete 2 Lecture Overviewrdquo httppeoplecegatechedusimkk92massconcretepdf
[3] H Weigler and S Karl Junger Beton Beanspruchung-Festigkeit-Verformung vol 40 Betonwerk Fertigteil-Technik 1974
[4] K A Riding J L Poole A K Schindler M C G Juenger andK J Folliard ldquoEvaluation of temperature prediction methodsformass concretemembersrdquoACIMaterials Journal vol 103 no5 pp 357ndash365 2006
[5] ConcreteWorks IHEEP Conference San Antonio Tex USA2009
[6] C T LeondesNeural Network Systems Techniques and Applica-tions Algorithms and Architectures Academic Press New YorkNY USA 1998
[7] R P Lippmann ldquoAn introduction to computing with neuralnetsrdquo IEEE ASSP Magazine vol 4 no 2 pp 4ndash22 1987
[8] D Sarkar ldquoMethods to speed up error back-propagation learn-ing algorithmrdquo ACMComputing Surveys vol 27 no 4 pp 519ndash542 1995
[9] F Ozcan CDAtis OKarahan EUncuoglu andHTanyildizildquoComparison of artificial neural network and fuzzy logic mod-els for prediction of long-term compressive strength of silicafume concreterdquo Advances in Engineering Software vol 40 no9 pp 856ndash863 2009
[10] H Demuth and M Beale Neural Network Toolbox for Use withMATLAB Handbook 2002
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 5
Difference between actual and ANN predicted temperature
Freq
uenc
y
200
150
100
50
0210minus1minus2minus3
Mean = 0Std Dev = 021119873 = 468
Figure 7 Distribution of the difference between predicted temper-ature by ANN and actual values in the test process
(iii) The inter correlation between input variables andconcrete temperature showed that higher environ-ment temperature and time in concrete can resultin higher concrete temperature and higher aggregatecontent in concrete results in lower-concrete temper-ature No other parameters were significant
(iv) The linear and nonlinear equations can estimate theconcrete temperature with correlation coefficients(1198772) of 0814 and 0873 respectively
(v) The artificial neural network procedure can predictthe concrete temperature with correlation coefficientof 0999 Therefore the obtained results are muchbetter than multivariate regression
(vi) The results show that artificial neural network isa reliable method to predict concrete temperatureduring curing
Conflict of Interests
Hereby the authors disclosure that this paper was just acontribution to the advancement of science and they justused SPSS and MATLAB softwares as the mathematicalmethods for prediction The authors do not have a directfinancial relation with the commercial identity mentioned inthe paper (SPSS and MATLAB softwares)
References
[1] P F Siew T Puapansawat and YHWu ldquoTemperature and heatstress in a concrete column at early agesrdquoANZIAM Journal vol44 no E pp C705ndashC722 2003
[2] R Moser ldquoMass Concrete CEE8813AmdashMaterial scienceof concrete 2 Lecture Overviewrdquo httppeoplecegatechedusimkk92massconcretepdf
[3] H Weigler and S Karl Junger Beton Beanspruchung-Festigkeit-Verformung vol 40 Betonwerk Fertigteil-Technik 1974
[4] K A Riding J L Poole A K Schindler M C G Juenger andK J Folliard ldquoEvaluation of temperature prediction methodsformass concretemembersrdquoACIMaterials Journal vol 103 no5 pp 357ndash365 2006
[5] ConcreteWorks IHEEP Conference San Antonio Tex USA2009
[6] C T LeondesNeural Network Systems Techniques and Applica-tions Algorithms and Architectures Academic Press New YorkNY USA 1998
[7] R P Lippmann ldquoAn introduction to computing with neuralnetsrdquo IEEE ASSP Magazine vol 4 no 2 pp 4ndash22 1987
[8] D Sarkar ldquoMethods to speed up error back-propagation learn-ing algorithmrdquo ACMComputing Surveys vol 27 no 4 pp 519ndash542 1995
[9] F Ozcan CDAtis OKarahan EUncuoglu andHTanyildizildquoComparison of artificial neural network and fuzzy logic mod-els for prediction of long-term compressive strength of silicafume concreterdquo Advances in Engineering Software vol 40 no9 pp 856ndash863 2009
[10] H Demuth and M Beale Neural Network Toolbox for Use withMATLAB Handbook 2002
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
