Main Article Content
Soft-computing techniques for fire safety parameter predictions in flammability studies are essential for describing a material fire behaviour. This study proposed, two novel Artificial Intelligence developed models, Multivariate Adaptive Regression Splines (MARS) and Random Forest (RF) methods, to model and predict peak heat release rate (pHRR) of Polymethyl methacrylate (PMMA) from Microscale Combustion Calorimetry (MCC) experiment. From the statistical analysis, MARS presented the highest coefficient of determination (R2) values of (0.9998) and (0.9996) for training and testing respectively, with low MAD, MAPE and RMSE values. Comparatively, MARS outperformed RF in the predictions of pHRR, through its model algorithms that generated optimized equations for pHRR predictions, covering all non-linearity points of the experimental data. Amongst the input variables (sample mass, THR, HRC, pTemp and pTime), heating rate (β), highly influenced pHRR outcome predictions from MARS and RF models. However, to validate the performance and applicability of the proposed models. Results of MARS and RF were benchmarked with that from Artificial Neural Network (ANN) methods. The MARS and RF models observed the least error deviation when compared with pHRR results for PMMA from the ANN models. This study therefore, recommends the adoption of MARS and RF in the predictions of flammability characteristics of polymeric materials.
(Accessed: 3 May 2019)
Matala A. Methods and applications of pyrolysis modeling for polymeric materials,
Doctor degree dissertation. VTT Science 44; 2013.
Lyon RE, Walters RN. Microscale Combustion Calorimeter, US Patent, US5981290A; 1999.
Xu Q, Jin C, Griffin GJ, Matala A, Hostikka S. A PMMA flammability analysis using the MCC:
Effect of specimen mass, J. Therm. Anal. Calorim. 2016;126:1831–1840.
Lyon RE. Heat release rate calorimeter for milligram samples, US Patent, US6464391B2; 2002.
Standard test method for determining flammability characteristics of plastics and other solid
materials using microscale combustion calorimetry ASTM D 7309-13, American Society for Testing
and Materials (International), West Conshohocken, PA; 2013.
Lyon RE, Walters RN, Stoliarov SI, and Safronava N. Principles and practice of micro-scale com
bustion calorimetry, Tech. Rep. DOT/FAA/TC- 12/53, US Department of Transportation, FAA; 2013.
Walters RN, Lyon RE. Microscale combustion calorimeter for determining flammability parameters
of materials. Evol Technol. Compet. Edge. 1997;42:1335–44.
Lyon RE, Walters RN, Stoliarov SI. Screening flame-retardants for plastics using Microscale Combustion
Calorimetry. Polym Eng. Sci. 2007;47(10):1501–10.
Xu Q, Jin C, Jiang Y. Compare the flammability of two extruded polystyrene foams with micro-scale
combustion calorimeter and cone calorimeter tests. Journal of Thermal Analysis and Calorimetry. 2017;127(3):2359-66.
Yang CQ, He Q, Lyon RE, Hu Y. Investigation of the flammability of different textile fabrics using micro-scale
combustion calorimetry. Polym Degrad Stab. 2010;95(2):108–15.
Xu Q, Jin C, Majlingova A, Restas A. Discuss the heat release capacity of polymer derived from microscale
combustion calorimeter. Journal of Therm. Anal. Calorim. 2017;133:1.
Xu Q, Jin C, Jiang Y. Analysis of the relationship between MCC and thermal analysis results in evaluating
flammability of EPS foam, J. Therm. Anal. Calorim. 2014;118(2):687–693.
Lyon RE, Takemori MT, Safronava N, Stoliarov SI, Walters RN. A molecular basis for polymer flammability,
Polymer 50. 2009;2608–2617.
Walters RN, Lyon RE. Molar group contributions to polymer flammability, J. Appl. Polym. Sci. 2003;87:548–563.
Parandekar PV, Browning AR, Prakash O. Modeling the flammability characteristics of polymers using
quantitative structure–property relationships (QSPR), Polym. Eng. Sci. 2015;55:1553–1559.
Asante-Okyere S, Xu Q, Mensah RA, Jin C, Ziggah YY. Generalized regression and feed forward back
propagation neural networks in modelling flammability characteristics of polymethyl methacrylate
(PMMA). Thermochimica Acta. 2018;667: 79-92.
Ma S, Lv M, Deng F, Zhang X, Zhai H, Lv W. Predicting the ecotoxicity of ionic liquids towards Vibrio
fischeri using Genetic Function Approximation and Least Squares Support Vector Machine, Journal of
Hazardous Materials. 2015;283:591-598.
Burgaz E, Yazici M, Kapusuz M, Alisir SH, Ozcan H. Prediction of thermal stability, crystallinity and
thermomechanical properties of poly(ethylene-oxide)/clay nanocomposites with Artificial Neural
Networks, Thermochi. Acta. 2014; 575:159-166.
Mensah RA, Jiang L, Asante-Okyere S, Xu Q, Jin C. Comparative evaluation of the predictability of
neural network methods on the flammability characteristics of extruded polystyrene from microscale
combustion calorimetry. J.Therm. Anal. Cal. 2019;133: 1–10.
Friedman JH. Multivariate adaptive regression splines. The Annals of Statistics. 1991;19(1):1-141.
Friedman JH, Roosen CB. An introduction to multivariate adaptive regression splines”, Stat Methods Med Res. 1995; 4:197.
Ziggah YY, Laari PB. Application of multivariate adaptive regression spline approach for 2D coordinate transformation.
Ghana Journal of Tech. 2018;2(2):50-62.
Xu QS, Massart DL, Liang YZ, Fang KT. Two-step multivariate adaptive regression splines for modeling a quantitative relationship between gas chromatography retention indices and molecular descriptors, Journal of Chromatography A. 2003;998(1–2):155-167.
Balshi MS, Mcguire AD, Duffy P, Flannigan M, Walsh J, Melillo J. Assessing the response of area burned to changing
climate in western boreal North America using a Multivariate Adaptive Regression Splines (MARS) approach. Global
Change Biology. 2009;15:578-600.
DOI:10.1111/j.1365-2486 .2008. 01679.x
Durmaz M, Karslioglu MO, Nohutcu M. Regional VTEC modeling with multivariate adaptive regression splines.
Advances in Space Research. 2010;46:180-189.
Roy SS, Roy R, Balas VE. Estimating heating load in buildings using multivariate adaptive regression splines, extreme learning machine, a hybrid model of MARS and ELM, Renew. Sustain. Energy Rev. 2018;82:4256–4268.
Breiman L. Random forests. Machine Learning. 2001;45:5–32.
Pramila PV, Mahesh V. Comparison of multivariate adaptive regression splines and random forest regression in
predicting forced expiratory volume in one second. International Journal of Bioengineering and Life Sciences. 2015;9 (4):338-342.
Palmer DS, O’Boyle NM, Glen CR, Mitchell JBO. Random forest models to predict aqueous solubility. J. of Chemical Inform. and Modeling. 2007;47(1):150-158.
Matin. SS, Chelgani SC. Estimation of coal gross calorific value based on various analyses by random forest method, Fuel. 2016;177:274-278.
Svetnik V, Liaw A, Tong CJ, Culberson C, Sheridan RP, and Feuston BP. Random forest: A classification and
regression tool for compound classification and QSAR modeling. Journal of Chemical Information and Computer Sciences. 2003;43(6): 1947-1958.
Govmark Datasheet of Micro-scale Combustion Calorimeter (MCC2), the Govmark Organization, Inc.
Bylander T. Estimating generalization error on two class datasets using out-of-bag estimates. Machine Learning. 2002; 48(18,22):287–297.
Liaw A, Wiener M. Classification and regression by random forest. R. News. 2002;2(3):18–22.
Nawar S, Mouazen M. Comparison between random forests, Artificial Neural Networks and Gradient Boosted Machines Methods of On-Line Vis-NIR Spectroscopy Measurements of Soil Total Nitrogen and Total Carbon, Sensors. 2017; 2428(17):1-22.
Segal M, Xiao Y. Multivariate random forests. WIREs Data Mining Knowledge Discovery. 2011;1:80–87.
Adoko AC, Jiao YY, Wu L, Wang H, Wang Z-H. Predicting tunnel convergence using multivariate adaptive regression spline and artificial neural network, tunneling and underground space tech. 2013;(38): 368-376.
Salford Predictive Modeler (SPM). RFTM, MARSTM User Guide. Salford Systems, San Diego, CA, USA; 2001.
Dey P, Das AK. Application of multivariate adaptive regression spline-assisted objective function on optimization of heat transfer rate around a cylinder, Nuclear Engineering and Tech. 2016;48:1315-1320.
Babrauskas V, Peacock RD. Heat release rate: the single most important variable in fire hazard. Fire Safety Journal. 1992;18: 255.