Optimization of micro-EDM parameters using grey-based fuzzy logic coupled with the Taguchi method Optimizacija parametrov mikroelektroerozije z uporabo mehke logike v povezavi s Taguchi metodo M. S. Vijayanand, M. Ilangkumaran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 989 Mechanism of multi-layer composite coatings in the zinc process of recycling coated WC-Co cemented-carbide scrap Mehanizem ve~plastnih kompozitnih premazov v procesu cinkanja za recikliranje odpadkov opla{~enih WC-Co karbidnih trdin H. Kuang, D. Tan, W. He, X. Wang, J. Zhong, H. Wang, C. Yang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 997 Basic physical, mechanical and electrical properties of electrically enhanced alkali-activated aluminosilicates Osnovne fizikalne, mehanske in elektri~ne lastnosti elektri~no izbolj{anih, z alkalijami aktiviranih aluminosilikatov L. Fiala, M. Jerman, P. Rovnaník, R. ^erný . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1005 Modeling of water removal in direct-chill casting of aluminum-alloy billets Modeliranje omejevanja neposrednega hlajenja z vodo med vertikalnim konti litjem gredic iz Al-zlitin A. Meysami, S. Mahmoudi, M. Hajisafari . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1011 Improving the microstructure and mechanical properties of magnesium-alloy sheets with a new extrusion method Izbolj{anje mikrostrukture in mehanskih lastnosti plo~evine iz Mg zlitine z novo metodo iztiskanja L. Lu, Z. Yin, Y. Liu, D. Chen, C. Liu, Z. Wu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019 Formation mechanism of diffusion-reaction layer for a Cu/Ti diffusion couple under different heating methods Oblikovanje mehanizma difuzijsko reakcijske plasti na Cu/Ti povr{ini z razli~nimi metodami segrevanja L. Fei, W. Mingfang, P. Juan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025 De-oxidation of PK942 steel with Ti and Zr Dezoksidacija jekla PK942 s Ti in Zr M. Kole`nik, J. Burja, B. [etina Bati~, A. Nagode, J. Medved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1031 Corrosion on polished and laser-textured surfaces of an Fe–Mn biodegradable alloy Primerjava korozijskih lastnosti polirane in lasersko teksturirane povr{ine biorazgradljive zlitine Fe–Mn M. Ho~evar, ^. Donik, I. Paulin, A. Kocijan, F. Tehovnik, J. Burja, P. Gregor~i~, M. Godec. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037 Comparison of the surface and anticorrosion properties of SiO2 and TiO2 nanoparticle epoxy coatings Primerjava povr{inskih in protikorozijskih lastnosti epoksidnih prevlek obogatenih s SiO2 in TiO2 nanovklju~ki M. Conradi, A. Kocijan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1043 LETNO KAZALO – INDEX Letnik 51 (2017), 1–6 – Volume 51 (2017), 1–6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1047 2 Materiali in tehnologije / Materials and technology 51 (2017) 6, MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 883–888 INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL USING GENETIC PROGRAMMING POVE^EVANJE NAPETOSTNE TRDNOSTI IN RAZTEZKA 16MnCrS5 JEKLA Z UPORABO GENETSKEGA PROGRAMIRANJA Miha Kova~i~1,2, Ana Turn{ek1, Darja Ocvirk1, Ga{per Gantar3 1[tore Steel d.o.o., @elezarska cesta 3, 3220 [tore, Slovenia 2Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia 3College of industrial Engineering, Mariborska cesta 2, 3000 Celje, Slovenia miha.kovacic@store-steel.si Prejem rokopisa – received: 2016-09-29; sprejem za objavo – accepted for publication: 2017-05-05 doi:10.17222/mit.2016.293 [tore Steel Ltd. is one of the largest spring-steel producers in Europe. [tore Steel makes more than 1400 steel grades of different chemical composition. Among them is 16MnCrS5 steel. It is generally used for the fabrication of case-hardened machine parts for various applications (e.g., bars, rods, plates, strips, forgings), where a combination of wear resistance, toughness and dynamic strength is essential. These properties can be easily correlated with tensile strength, which depends on the chemical composition and heat treatment after rolling. In addition, the elongation should be taken into account. In the paper, modeling of tensile strength and elongation with genetic programming is presented and compared with linear regression modeling. The chemical composition data (content of C, Mn, S and Cr) and the heat-treatment regime data (GKZ or BG annealing) were used for modeling. The modeling results show that a higher tensile strength with improved elongation were achieved. Keywords: 16MnCrS5, tensile strength, elongation, modeling, genetic programming, linear regression [tore Steel je najve~ji proizvajalec vzmetnega jekla v Evropi. [tore Steel izdeluje ve~ kot 1400 razli~nih kvalitet jekla z razli~nimi kemijskimi sestavami. Med njmi tudi 16MnCrS5, ki spada v skupino jekel za cementacijo, ki so namenjena za strojno obdelavo razli~nih delov (npr. palic, plo{~, trakov, odkovkov), kjer se zahteva kombinacija obrabne odpornosti, `ilavosti ter trajnonihajne trdnosti. Le-te lastnosti lahko povezujemo z natezno trdnostjo, ki je odvisna predvsem od kemi~ne sestave in toplotne obdelave po valjanju. Prav tako je pomemben raztezek. V ~lanku je predstavljeno modeliranje natezne trdnosti in raztezka s pomo~jo genetskega modeliranja in linearne regresije. Za modeliranje smo uporabili vsebnosti kemijskih elementov (C, Mn, S in Cr) ter na~in toplotne obdelave (GKZ ali BG). Glede na rezultate smo pove~ali natezno trdnost pri izbolj{anem raztezku. Klju~ne besede: 16MnCrS5, natezna trdnost, raztezek, modeliranje, genetsko programiranje, linearna regresija 1 INTRODUCTION In the modern steel-production and steel-consump- tion industry, it is essential to know the material pro- perties and behavior. There are several well-known commercial types of software available for modeling material properties but steel producers are often forced on using inventive experiments, methods and approaches due to their unique production, equipment, time con- straints and niche applications.1–6 The literature review reveals that tensile strength and elongation optimization of steel products in general incorporates multi-criteria optimization approaches,1,2,7–9 based also on artificial intelligence methods.10–14This is the case for rated material properties in both quantitative and qualitative terms.1,10 The required tensile strength and elongation are ob- tained by changing: • chemical composition,2,12,14 • plastic deformation parameters (i.e., influencing mi- crostructure, grain size),8,15–17 • heat-treatment parameters after plastic deforma- tion.2,7,14 The article presents the practical implementation of tensile strength and elongation optimization for long- rolled products made of 16MnCrS5 steel, which is generally used for the fabrication of case-hardened machine parts for several applications (e.g., bars, rods, plates, strips, forgings). First, we provide the experimental background, then the methods used, and, lastly, we present the practical implementation and draw conclusions. 2 EXPERIMENTAL BACKGROUND In general, production starts with scrap melting in an electro arc furnace (EAF).18 After the scrap and car- burizing agents have been melted, dephosphorization is conducted. The melting bath is heated up to the tapping temperature and, after secondary steel treatment, is discharged into the casting ladle. After discharging from EAF, the melting bath is deoxidated, desulphurized, the nonmetallic inclusions Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 883 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS UDK 620.172.2:621.3.015:519.7 ISSN 1580-2949 Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 51(6)883(2017) are filtered out, the slag metallic oxides are reduced, the hydrogen and nitrogen are partly degassed, the melting bath and temperature field are homogenized, and then the formed slag exchange and the major alloying are carried out. The melt is poured from the casting ladle into the tundish on the continuous casting device, where 180 mm square billets from 2 m to 4 m in length are cast. All billets are cooled before the rolling operation. The billets are heated, in accordance with the pre- scribed temperatures, in the continuous heating furnace. After heating, the billets are hot rolled on the rolling stands. Depending on the various shapes and dimensions of the grooves, cylindrical, square or flat steel bars can be produced. Steel bars are cooled on the cooling bed. After cooling they are cut into different lengths using hot shears. During cutting, the samples for examination of the material are taken (e.g., tensile strength, hardenabi- lity, micro-cleanliness). After cooling, the bars can be additionally heat treated in accordance with the customers’ orders. In [tore Steel there are 7 different quality pres- criptions (chemical compositions) for 16MnCrS5.19 Only the most representative one (Table 1) was used in the research. From October 2008 to February 2016, 70 consecutively cast batches (orders) with diameters from 20 mm to 55 mm were produced. In all cases the data on tensile strength and elongation is available. For some orders the bars were also heat treated: • GKZ (Glühen auf kugeligen Zementit) – spheroidiza- tion annealing or • BG annealing – isothermal annealing (for excellent machinability properties and a better micro-structure homogenizing). • The heat treatment is conducted using the tunnel heat-treatment furnace.3 Table 1: Chemical composition in mass fractions (w/%) of the most representative quality prescription (17) for 16MnCrS5 C Mn Cr S Mo Ni Mini- mum 0.14 1.00 0.8 0.020 Maxi- mum 0.19 1.30 1.1 0.040 0.08 0.30 During the microstructure examination, also the pear- lite content and percentage of spheroidization (after GKZ annealing) were assessed. The micro-cleanliness (K3) was determined according to DIN 50602, method K. All the metallurgical examinations and tensile testing were conducted by the head of the metallurgical labora- tory and by the person responsible for mechanical testing. The collected data is presented in Table 2. In the same table, besides quantitative (i.e., chemical compo- sition, content of pearlite, percentage of spheroidization and micro-cleanliness), also the qualitative parameter is included (heat treatment), where the prediction of material properties is often aggravated.1,10 Table 2: The collected data on 16MnCrS5 (quality prescription 17) B at ch # C (% ) M n (% ) S (% ) C r (% ) K 3 P ea rl it e (% ) S ph er od iz at i on (% ) H ea t tr ea tm en t R m (N /m m 2 ) A (% ) 1 0.17 1.26 0.03 1 7 40 0 / 669 24.5 2 0.17 1.22 0.027 0.99 5 40 0 / 604 17.2 3 0.17 1.22 0.027 0.99 6 45 0 / 658 16.1 4 0.18 1.2 0.025 0.97 6 40 0 / 654 16.2 5 0.17 1.19 0.023 0.99 7 45 0 / 666 16.8 6 0.19 1.29 0.023 1.09 9 40 90 GKZ 437 29.5 7 0.15 1.05 0.033 0.82 3 45 80 GKZ 454 30.4 8 0.15 1.02 0.028 0.82 11 40 90 GKZ 440 28.9 9 0.19 1.3 0.019 1.06 9 40 95 GKZ 446 30.5 10 0.19 1.27 0.021 1.07 15 40 0 BG 570 22.2 11 0.15 1.02 0.03 0.85 7 40 0 BG 593 22.3 … … … … … … … … … … … 70 0.16 1.09 0.029 0.92 5 40 0 / 644 15.2 3 MODELING OF TENSILE STRENGTH AND ELONGATION In Table 2 the percentage of spheroidization depends on GKZ (Glühen auf kugeligen Zementit) – spheroidi- zation annealing. Accordingly, both "qualitative" para- meters of GKZ and BG annealing can be replaced, respectively, by the quantitative values 0 and 1 (0 for when no heat treatment is used; 1 for when the heat treatment is used). For the model fitness the average relative deviation between the predicted and the experimental data was selected. It is defined as Equation (1): Δ = − = ∑ E P n i i i n 1 (1) where n is the size of the collected data, and Ei and Pi are the actual and the predicted total tensile strength or elongation, respectively. The tensile strength and elongation were modeled using linear regression and genetic programming. The results of the modeling are presented hereafter. 3.1 Linear regression Based on the linear regression results regarding tensile strength only the percent of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experimental data is 3.06 %. The linear-regression model for the tensile- strength prediction is in Equation (2): Rm = 500.0126·C – 198.777·Mn + 119.686·S + 128.071·Cr – 0.0738·K3 + 1.815·Pearlite – 1.66·Spherodization + 587.2196 (2) M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 884 Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS During the elongation modeling only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experi- mental data is 7.56 %. The linear regression model for the tensile-strength prediction is: A = –57.295·C + 21.749·Mn ± 44.973·S – 12.498·Cr – 0.0093·K3 – 0.0247·Pearliten + 0.07823·Spherodization + 16.718 (3) It must be emphasized that the percentage of sphe- roidization and heat treatment statistically significantly influences both the tensile strength and the elongation. However, it does so conversely (i.e., if tensile strength is increased by the influence of both parameters, the elongation is decreased and vice versa). Figure 1 shows the calculated influences of indivi- dual parameters on the tensile strength and elongation using the developed models, while separately changing the individual parameter within the range from Table 1. It can be concluded that the C, Mn, S and Cr content, the percentage of spheroidization and the heat treatment are the most influential factors. 3.2 Genetic programming Genetic programming has already been found useful for several different applications in [tore Steel Ltd.3,6,18,19 Genetic programming is a population-based algorithm that is similar to a genetic algorithm and many other heuristic optimization techniques.20–23 In the present paper 100 models for tensile strength and also for elon- gation were obtained through a genetic programming method. During the simulated evolution the organisms (with basic ingredients – function and terminal genes) are generated and afterwards changed through different changing algorithms. The following function genes were selected: addition (+), subtraction (-), multiplication (*) and division (/). The selected terminal genes were: weight percentage of C (C), Mn (MN), S (S), Cr (CR), the micro-cleanliness K3 (K3) determined according to DIN 50602, method K, content of pearlite in % (PEARLITE), percentage of spheroidization (SPHER) and heat treatment (HT). The AutoLISP-based in-house genetic programming system was run 100 times in order to develop 100 independent civilizations. Each run lasted approximately 4 h and 40 min on a 3.0-GHz processor with 4 GB of RAM. M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 885 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS Figure 1: Calculated influences of individual parameters on the tensile strength and elongation, while separately changing them within the range from Table 2 Figure 2: Calculated influences of individual parameters on the ten- sile strength and elongation, while separately changing them within the range from Table 2 (4) are filtered out, the slag metallic oxides are reduced, the hydrogen and nitrogen are partly degassed, the melting bath and temperature field are homogenized, and then the formed slag exchange and the major alloying are carried out. The melt is poured from the casting ladle into the tundish on the continuous casting device, where 180 mm square billets from 2 m to 4 m in length are cast. All billets are cooled before the rolling operation. The billets are heated, in accordance with the pre- scribed temperatures, in the continuous heating furnace. After heating, the billets are hot rolled on the rolling stands. Depending on the various shapes and dimensions of the grooves, cylindrical, square or flat steel bars can be produced. Steel bars are cooled on the cooling bed. After cooling they are cut into different lengths using hot shears. During cutting, the samples for examination of the material are taken (e.g., tensile strength, hardenabi- lity, micro-cleanliness). After cooling, the bars can be additionally heat treated in accordance with the customers’ orders. In [tore Steel there are 7 different quality pres- criptions (chemical compositions) for 16MnCrS5.19 Only the most representative one (Table 1) was used in the research. From October 2008 to February 2016, 70 consecutively cast batches (orders) with diameters from 20 mm to 55 mm were produced. In all cases the data on tensile strength and elongation is available. For some orders the bars were also heat treated: • GKZ (Glühen auf kugeligen Zementit) – spheroidiza- tion annealing or • BG annealing – isothermal annealing (for excellent machinability properties and a better micro-structure homogenizing). • The heat treatment is conducted using the tunnel heat-treatment furnace.3 Table 1: Chemical composition in mass fractions (w/%) of the most representative quality prescription (17) for 16MnCrS5 C Mn Cr S Mo Ni Mini- mum 0.14 1.00 0.8 0.020 Maxi- mum 0.19 1.30 1.1 0.040 0.08 0.30 During the microstructure examination, also the pear- lite content and percentage of spheroidization (after GKZ annealing) were assessed. The micro-cleanliness (K3) was determined according to DIN 50602, method K. All the metallurgical examinations and tensile testing were conducted by the head of the metallurgical labora- tory and by the person responsible for mechanical testing. The collected data is presented in Table 2. In the same table, besides quantitative (i.e., chemical compo- sition, content of pearlite, percentage of spheroidization and micro-cleanliness), also the qualitative parameter is included (heat treatment), where the prediction of material properties is often aggravated.1,10 Table 2: The collected data on 16MnCrS5 (quality prescription 17) B at ch # C (% ) M n (% ) S (% ) C r (% ) K 3 P ea rl it e (% ) S ph er od iz at i on (% ) H ea t tr ea tm en t R m (N /m m 2 ) A (% ) 1 0.17 1.26 0.03 1 7 40 0 / 669 24.5 2 0.17 1.22 0.027 0.99 5 40 0 / 604 17.2 3 0.17 1.22 0.027 0.99 6 45 0 / 658 16.1 4 0.18 1.2 0.025 0.97 6 40 0 / 654 16.2 5 0.17 1.19 0.023 0.99 7 45 0 / 666 16.8 6 0.19 1.29 0.023 1.09 9 40 90 GKZ 437 29.5 7 0.15 1.05 0.033 0.82 3 45 80 GKZ 454 30.4 8 0.15 1.02 0.028 0.82 11 40 90 GKZ 440 28.9 9 0.19 1.3 0.019 1.06 9 40 95 GKZ 446 30.5 10 0.19 1.27 0.021 1.07 15 40 0 BG 570 22.2 11 0.15 1.02 0.03 0.85 7 40 0 BG 593 22.3 … … … … … … … … … … … 70 0.16 1.09 0.029 0.92 5 40 0 / 644 15.2 3 MODELING OF TENSILE STRENGTH AND ELONGATION In Table 2 the percentage of spheroidization depends on GKZ (Glühen auf kugeligen Zementit) – spheroidi- zation annealing. Accordingly, both "qualitative" para- meters of GKZ and BG annealing can be replaced, respectively, by the quantitative values 0 and 1 (0 for when no heat treatment is used; 1 for when the heat treatment is used). For the model fitness the average relative deviation between the predicted and the experimental data was selected. It is defined as Equation (1): Δ = − = ∑ E P n i i i n 1 (1) where n is the size of the collected data, and Ei and Pi are the actual and the predicted total tensile strength or elongation, respectively. The tensile strength and elongation were modeled using linear regression and genetic programming. The results of the modeling are presented hereafter. 3.1 Linear regression Based on the linear regression results regarding tensile strength only the percent of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experimental data is 3.06 %. The linear-regression model for the tensile- strength prediction is in Equation (2): Rm = 500.0126·C – 198.777·Mn + 119.686·S + 128.071·Cr – 0.0738·K3 + 1.815·Pearlite – 1.66·Spherodization + 587.2196 (2) M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 884 Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS During the elongation modeling only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experi- mental data is 7.56 %. The linear regression model for the tensile-strength prediction is: A = –57.295·C + 21.749·Mn ± 44.973·S – 12.498·Cr – 0.0093·K3 – 0.0247·Pearliten + 0.07823·Spherodization + 16.718 (3) It must be emphasized that the percentage of sphe- roidization and heat treatment statistically significantly influences both the tensile strength and the elongation. However, it does so conversely (i.e., if tensile strength is increased by the influence of both parameters, the elongation is decreased and vice versa). Figure 1 shows the calculated influences of indivi- dual parameters on the tensile strength and elongation using the developed models, while separately changing the individual parameter within the range from Table 1. It can be concluded that the C, Mn, S and Cr content, the percentage of spheroidization and the heat treatment are the most influential factors. 3.2 Genetic programming Genetic programming has already been found useful for several different applications in [tore Steel Ltd.3,6,18,19 Genetic programming is a population-based algorithm that is similar to a genetic algorithm and many other heuristic optimization techniques.20–23 In the present paper 100 models for tensile strength and also for elon- gation were obtained through a genetic programming method. During the simulated evolution the organisms (with basic ingredients – function and terminal genes) are generated and afterwards changed through different changing algorithms. The following function genes were selected: addition (+), subtraction (-), multiplication (*) and division (/). The selected terminal genes were: weight percentage of C (C), Mn (MN), S (S), Cr (CR), the micro-cleanliness K3 (K3) determined according to DIN 50602, method K, content of pearlite in % (PEARLITE), percentage of spheroidization (SPHER) and heat treatment (HT). The AutoLISP-based in-house genetic programming system was run 100 times in order to develop 100 independent civilizations. Each run lasted approximately 4 h and 40 min on a 3.0-GHz processor with 4 GB of RAM. M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 885 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS Figure 1: Calculated influences of individual parameters on the tensile strength and elongation, while separately changing them within the range from Table 2 Figure 2: Calculated influences of individual parameters on the ten- sile strength and elongation, while separately changing them within the range from Table 2 (4) The selected maximum number of generations was 100. The selected size of the population of organisms 100, the reproduction probability 0.4, the crossover pro- bability 0.6, the maximum permissible depth in the creation of the population 6, the maximum permissible depth after the operation of crossover 10, and the smallest permissible depth of the organisms in gene- rating new organisms 2. For the selection of organisms the tournament method with tournament size 7 was used. For the tensile-strength modeling, the most succesful organism from all of the civilizations is presented in Equation (4). The average relative deviation between the predicted and the experimental data is 2.89 %. The model consists of 235 genes. Its depth is 14. It must also be emphasized that the parameters (genes) CR, S, K3 are not included in the model. For elongation modeling, the most succesful organism from all of the civilizations is presented in Equation (5). The average relative deviation between the predicted and the experimental data is 5.18 %. The model consists of 189 genes. Its depth is 14. It must also be emphasized that parameters (genes) CR, S, K3 are not included in the model. Figure 2 shows the calculated influences of the indi- vidual parameters on the tensile strength and elongation using the genetically developed model (Equations (4) and (5)), while separately changing the individual para- meter within the range from Table 2. It can be concluded that Cr and pearlite content, the percentage of sphe- roidization and the heat treatment are the most influential factors. 4 IMPLEMENTATION OF MODELING RESULTS Up until February 2016, six consecutively cast batches (16MnCrS5, quality prescription 17) were used for the implementation of the modeling results. The collected data are presented in Table 3. Depending on the chemical composition and the pearlite content, the plan for the heat treatment is determined in order to achieve the optimal tensile strength at moderate elon- gation. Table 3: The collected data on 6 16MnCrS5 (quality prescription 17) consecutively cast batches Batch # C (%) Mn (%) S (%) Cr (%) K3 Pearlite (%) 1 0.17 1.09 0.035 0.87 4 40 2 0.17 1.13 0.025 0.86 8 45 3 0.16 1.08 0.026 0.85 6 40 4 0.18 1.09 0.022 0.85 7 40 5 0.15 1.07 0.023 0.87 6 40 6 0.17 1.13 0.025 0.99 7 40 Figure 3 presents the prediction of tensile strength using linear regression and genetic programming for the 6 batches (from Table 3). There are statistically signi- ficant differences (one-way ANOVA, p < 0.05) between the predicted tensile strength when the heat treatment is used and when it is not used – for both methods: linear regression and genetic programming. GKZ annealed material has the statistically significantly lowest tensile strength (one-way ANOVA, p < 0.05). Figure 4 presents the prediction of elongation using linear regression and genetic programming for 6 batches (from Table 3). There are statistically significant diffe- rences (one-way ANOVA, p < 0.05) between the predic- ted elongation when the heat treatment is used and when the heat treatment is not used – for both methods: linear regression and genetic programming. GKZ annealed ma- M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 886 Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS (5) Figure 3: Calculated influences of the individual parameters on the elongation, while separately changing them within the range from Table 2 terial has the statistically significantly highest elongation (one-way ANOVA, p < 0.05). In accordance with the predictions presented in Fig- ures 3 and 4 the BG annealing was selected in order to achieve a higher tensile strength at moderate elongation. The tensile-test results (tensile strength and elonga- tion) and linear-regression results are presented in Table 4. The average relative deviation between the predicted and the experimental data for the tensile strength for linear regression and genetic programming is 3.55 % and 3.08 %, respectively. The average relative deviation bet- ween the predicted and the experimental data for elon- gation for linear regression and genetic programming is 3.56 % and 2.65 %, respectively. 5 CONCLUSIONS The 16MnCrS5 steel grade is generally used for the fabrication of case-hardened machine parts for several applications (e.g., bars, rods, plates, strips, forgings), where having a combination of wear resistance, tough- ness and dynamic strength is essential. These qualities can be easily correlated with tensile test results (e.g., ten- sile strength, elongation), which depends on the chemi- cal composition and the heat treatment after rolling. Accordingly, from October 2008 to February 2016, 70 consecutively cast batches with diameters from 20 mm to 55 mm were produced. In all cases the data on the tensile strength and elongation is available. For some orders the bars were also heat treated: • GKZ (Glühen auf kugeligen Zementit) – spheroidi- zation annealing or • BG annealing – isothermal annealing (for excellent machinability properties and a better microstructure homogenizing). The chemical composition (content of C, Mn, S and Cr) and heat-treatment regime (GKZ or BG annealing) data were collected. During the microstructure examina- tion also the pearlite content and the percentage of spheroidization (after GKZ annealing) were assessed. The micro-cleanliness (K3) was determined according to DIN 50602, method K. The tensile strength and elongation were modeled using linear regression and genetic programming. The ANOVA results obtained using linear regression show that only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05) for tensile strength. The average relative deviation between the predicted and the experi- mental data is 3.06 %. At elongation only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experimental data is 7.56 %. The average relative deviation between the predicted and the experimental data for the best genetically deve- loped mathematical model for predicting tensile strength is 2.89 %. The model consists of 235 genes. Its depth is 14. The average relative deviation between the predicted and the experimental data for the best genetically deve- loped mathematical model for predicting elongation is 5.18 %. The model consists of 189 genes. Its depth is 14. Up until February 2016, six consecutively cast batches (16MnCrS5, quality prescription 17) were used for the implementation of the modeling results. To achieve the optimal tensile strength at moderate elonga- tion, the plan for the heat treatment has to be determined on the basis of the chemical composition and the pearlite content. The tensile strength and the elongation were predicted using linear regression and genetic programm- M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 887 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS Figure 4: Calculated influences of the individual parameters on the elongation, while separately changing them within the range from Table 2 Table 4: The collected data on 6 16MnCrS5 (quality prescription 17) consecutively cast batches B at ch # C (% ) M n (% ) S (% ) C r (% ) K 3 P ea rl it e (% ) R m (M Pa ) R m – li ne ar re gr es si on (M Pa ) R m – G P (M Pa ) A (% ) A – li ne ar re gr es si on (% ) A – G P (% ) 1 0.17 1.09 0.035 0.87 4 40 622.0 600.28 590.53 22.60 21.82 16.85 2 0.17 1.13 0.025 0.86 8 45 641.0 598.63 599.92 22.00 23.10 14.02 3 0.16 1.08 0.026 0.85 6 40 578.0 593.48 590.74 21.90 22.81 16.91 4 0.18 1.09 0.022 0.85 7 40 613.0 600.94 590.54 23.00 22.05 17.10 5 0.15 1.07 0.023 0.87 6 40 607.0 592.67 590.98 22.70 23.05 16.89 6 0.17 1.13 0.025 0.99 7 40 598.0 606.28 589.95 22.30 21.61 17.05 The selected maximum number of generations was 100. The selected size of the population of organisms 100, the reproduction probability 0.4, the crossover pro- bability 0.6, the maximum permissible depth in the creation of the population 6, the maximum permissible depth after the operation of crossover 10, and the smallest permissible depth of the organisms in gene- rating new organisms 2. For the selection of organisms the tournament method with tournament size 7 was used. For the tensile-strength modeling, the most succesful organism from all of the civilizations is presented in Equation (4). The average relative deviation between the predicted and the experimental data is 2.89 %. The model consists of 235 genes. Its depth is 14. It must also be emphasized that the parameters (genes) CR, S, K3 are not included in the model. For elongation modeling, the most succesful organism from all of the civilizations is presented in Equation (5). The average relative deviation between the predicted and the experimental data is 5.18 %. The model consists of 189 genes. Its depth is 14. It must also be emphasized that parameters (genes) CR, S, K3 are not included in the model. Figure 2 shows the calculated influences of the indi- vidual parameters on the tensile strength and elongation using the genetically developed model (Equations (4) and (5)), while separately changing the individual para- meter within the range from Table 2. It can be concluded that Cr and pearlite content, the percentage of sphe- roidization and the heat treatment are the most influential factors. 4 IMPLEMENTATION OF MODELING RESULTS Up until February 2016, six consecutively cast batches (16MnCrS5, quality prescription 17) were used for the implementation of the modeling results. The collected data are presented in Table 3. Depending on the chemical composition and the pearlite content, the plan for the heat treatment is determined in order to achieve the optimal tensile strength at moderate elon- gation. Table 3: The collected data on 6 16MnCrS5 (quality prescription 17) consecutively cast batches Batch # C (%) Mn (%) S (%) Cr (%) K3 Pearlite (%) 1 0.17 1.09 0.035 0.87 4 40 2 0.17 1.13 0.025 0.86 8 45 3 0.16 1.08 0.026 0.85 6 40 4 0.18 1.09 0.022 0.85 7 40 5 0.15 1.07 0.023 0.87 6 40 6 0.17 1.13 0.025 0.99 7 40 Figure 3 presents the prediction of tensile strength using linear regression and genetic programming for the 6 batches (from Table 3). There are statistically signi- ficant differences (one-way ANOVA, p < 0.05) between the predicted tensile strength when the heat treatment is used and when it is not used – for both methods: linear regression and genetic programming. GKZ annealed material has the statistically significantly lowest tensile strength (one-way ANOVA, p < 0.05). Figure 4 presents the prediction of elongation using linear regression and genetic programming for 6 batches (from Table 3). There are statistically significant diffe- rences (one-way ANOVA, p < 0.05) between the predic- ted elongation when the heat treatment is used and when the heat treatment is not used – for both methods: linear regression and genetic programming. GKZ annealed ma- M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 886 Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS (5) Figure 3: Calculated influences of the individual parameters on the elongation, while separately changing them within the range from Table 2 terial has the statistically significantly highest elongation (one-way ANOVA, p < 0.05). In accordance with the predictions presented in Fig- ures 3 and 4 the BG annealing was selected in order to achieve a higher tensile strength at moderate elongation. The tensile-test results (tensile strength and elonga- tion) and linear-regression results are presented in Table 4. The average relative deviation between the predicted and the experimental data for the tensile strength for linear regression and genetic programming is 3.55 % and 3.08 %, respectively. The average relative deviation bet- ween the predicted and the experimental data for elon- gation for linear regression and genetic programming is 3.56 % and 2.65 %, respectively. 5 CONCLUSIONS The 16MnCrS5 steel grade is generally used for the fabrication of case-hardened machine parts for several applications (e.g., bars, rods, plates, strips, forgings), where having a combination of wear resistance, tough- ness and dynamic strength is essential. These qualities can be easily correlated with tensile test results (e.g., ten- sile strength, elongation), which depends on the chemi- cal composition and the heat treatment after rolling. Accordingly, from October 2008 to February 2016, 70 consecutively cast batches with diameters from 20 mm to 55 mm were produced. In all cases the data on the tensile strength and elongation is available. For some orders the bars were also heat treated: • GKZ (Glühen auf kugeligen Zementit) – spheroidi- zation annealing or • BG annealing – isothermal annealing (for excellent machinability properties and a better microstructure homogenizing). The chemical composition (content of C, Mn, S and Cr) and heat-treatment regime (GKZ or BG annealing) data were collected. During the microstructure examina- tion also the pearlite content and the percentage of spheroidization (after GKZ annealing) were assessed. The micro-cleanliness (K3) was determined according to DIN 50602, method K. The tensile strength and elongation were modeled using linear regression and genetic programming. The ANOVA results obtained using linear regression show that only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05) for tensile strength. The average relative deviation between the predicted and the experi- mental data is 3.06 %. At elongation only the percentage of spheroidization and heat treatment are statistically significant influential parameters (p < 0.05). The average relative deviation between the predicted and the experimental data is 7.56 %. The average relative deviation between the predicted and the experimental data for the best genetically deve- loped mathematical model for predicting tensile strength is 2.89 %. The model consists of 235 genes. Its depth is 14. The average relative deviation between the predicted and the experimental data for the best genetically deve- loped mathematical model for predicting elongation is 5.18 %. The model consists of 189 genes. Its depth is 14. Up until February 2016, six consecutively cast batches (16MnCrS5, quality prescription 17) were used for the implementation of the modeling results. To achieve the optimal tensile strength at moderate elonga- tion, the plan for the heat treatment has to be determined on the basis of the chemical composition and the pearlite content. The tensile strength and the elongation were predicted using linear regression and genetic programm- M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 887 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS Figure 4: Calculated influences of the individual parameters on the elongation, while separately changing them within the range from Table 2 Table 4: The collected data on 6 16MnCrS5 (quality prescription 17) consecutively cast batches B at ch # C (% ) M n (% ) S (% ) C r (% ) K 3 P ea rl it e (% ) R m (M Pa ) R m – li ne ar re gr es si on (M Pa ) R m – G P (M Pa ) A (% ) A – li ne ar re gr es si on (% ) A – G P (% ) 1 0.17 1.09 0.035 0.87 4 40 622.0 600.28 590.53 22.60 21.82 16.85 2 0.17 1.13 0.025 0.86 8 45 641.0 598.63 599.92 22.00 23.10 14.02 3 0.16 1.08 0.026 0.85 6 40 578.0 593.48 590.74 21.90 22.81 16.91 4 0.18 1.09 0.022 0.85 7 40 613.0 600.94 590.54 23.00 22.05 17.10 5 0.15 1.07 0.023 0.87 6 40 607.0 592.67 590.98 22.70 23.05 16.89 6 0.17 1.13 0.025 0.99 7 40 598.0 606.28 589.95 22.30 21.61 17.05 ing. Based on the obtained results the BG annealing was selected in order to achieve a higher tensile strength at moderate elongation. The average relative deviation between the predicted and the experimental data for the tensile strength for linear regression and genetic programming is 3.55 % and 3.08 %, respectively. The average relative deviation bet- ween the predicted and the experimental data for elon- gation for linear regression and genetic programming is 3.56 % and 2.65 %, respectively. In the future the proposed concept will be used for analyzing the properties of selected steel grades. 6 REFERENCES 1 S. Bruschi, T. Altan, D. Banabic, P.F.Bariani, A. Brosius, J. Cao, A. Ghiotti, M. Khraisheh, M. Merklein, A. E. Tekkaya, Testing and modelling of material behaviour and formability in sheet metal forming, CIRP Annals – Manufacturing Technology, 63 (2014) 2, 727–749, doi:10.1016/j.cirp.2014.05.005 2 Q. Zhang, M. Mahfouf, A nature-inspired multi-objective optimi- sation strategy based on a new reduced space searching algorithm for the design of alloy steels, Engineering Applications of Artificial Intelligence, 23 (2010) 5, 660–675, doi:10.1016/j.engappai.2010. 01.017 3 M. Kova~i~, B. [arler, Application of the Genetic Programming for Increasing the Soft Annealing Productivity in Steel Industry, Materials and Manufacturing Processes, 24 (2009) 3, 369–374, doi:10.1080/10426910802679634 4 M. Kova~i~, Modeling of Total Decarburization of Spring Steel with Genetic Programming, Materials and Manufacturing Processes, 30 (2014) 4, 434–443, doi:10.1080/10426914.2014.961477 5 M. Kova~i~, D. Novak, Prediction of the chemical non-homogeneity of 30MnVS6 billets with genetic programming, Materiali in tehnologije, 50 (2016) 1, 69–74, doi:10.17222/mit.2014.280 6 M. Kova~i~, R. Jager, Modeling of occurrence of surface defects of C45 steel with genetic programming, Materiali in tehnologije, 49 (2015) 6, 857–863, doi:10.17222/mit.2013.304 7 Z. Wang, P. Liu, Y. Xu, Y. Wang, Y. Zhan, Hot Stamping of High Strength Steel with Tailored Properties by Two Methods, Procedia Engineering, 81 (2014), 1725–1730, doi:10.1016/j.proeng.2014. 10.221 8 A. K. Lakshminarayanan, Enhancing the properties of friction stir welded stainless steel joints via multi-criteria optimization, Archives of Civil and Mechanical Engineering, 16 (2016) 4, 605–617, doi:10.1016/j.acme.2016.03.012 9 X. Kong, L. Lan, Z. Hu, B. Li, T. Sui, Optimization of mechanical properties of high strength bainitic steel using thermo-mechanical control and accelerated cooling process, Journal of Materials Pro- cessing Technology, 217 (2015), 202–210, doi:10.1016/j.jmatprotec. 2014.11.016 10 R. S. Khabba, B. D. Manshadi, A. Abedian, R. Mahmudi, A sim- plified fuzzy logic approach for materials selection in mechanical engineering design, Materials and Design, 30 (2009) 3, 687–697, doi:10.1016/j.matdes.2008.05.026 11 D. Pandya, D. Shah, Experimentation and its Prediction of Process Parameters Effects on Elongation in Tensile Test of AISI 1008 Steel Using ANN Model, Procedia Technology, 14 (2014), 282–289, doi:10.1016/j.protcy.2014.08.037 12 I. Mohanty, D. Bhattacharjee, S. Datta, Designing cold rolled IF steel sheets with optimized tensile properties using ANN and GA, Computational Materials Science, 50 (2011) 8, 2331–2337, doi:10.1016/j.commatsci.2011.03.007 13 Y. Y. Yang, M. Mahfouf, D. A. Linkens, Q. Zhang, Tensile Strength Prediction for Hot Rolled Steels by Bayesian Neural Network Model, IFAC Proceedings Volumes, Vol. 42 (2009) 23, 255–260, doi:10.3182/20091014-3-CL-4011.00046 14 J. Chen, M. Mahfouf, G. Sidahmed, A new holistic systems approach to the design of heat treated alloy steels using a biologically inspired multi-objective optimisation algorithm, Engineering Applications of Artificial Intelligence, 37 (2015), 103–114, doi:10.1016/j.engappai. 2014.08.014 15 J. Han, H. Li, Z. Zhu, L. Jiang, H. Xu, L. Ma, Effects of processing optimisation on microstructure, texture, grain boundary and mechanical properties of Fe-17Cr ferritic stainless steel thick plates, Materials Science and Engineering: A, 616 (2014), 20–28, doi:10.1016/j.msea.2014.07.107 16 S. Chen, X. Jin, L. Rong, Improving the strength and ductility of reduced activation ferritic/martensitic steel by cold-swaging and post-annealing, Materials Science and Engineering: A, 631 (2015), 139–143, doi:10.1016/j.msea.2015.02.044 17 L. Chen, J, Huang, Y. Zhao, H. Di, F. Zhu, Processing, Microstruc- tures and Mechanical Properties of Ultra-high Strength Steel Sheet, Procedia Engineering, 81 (2014), 84–89, doi:10.1016/j.proeng. 2014.09.131 18 S. Klemen, M. Kova~i~, P. Kitak, J. Pihler, Electric-arc-furnace pro- ductivity optimization, Materiali in tehnologije, 48 (2014) 1, 3–7 19 M. Kova~i~, B. [arler, Genetic Algorithm-Based Batch Filling Sche- duling in the Steel Industry, Materials and manufacturing processes, 26 (2011) 3, 464–474, doi:10.1080/10426914.2010.525576 20 J. F. Wang, W. L. Kang, J. L. Zhao, K. Y. Chu, A simulation approach to the process planning problem using a modified particle swarm optimization, Advances in Production Engineering & Management, 11 (2016) 2, 77–92, doi:10.14743/apem2016.2.211 21 S. Ramadan, A bi-objective inspection policy optimization model for finite-life repairable systems using a genetic algorithm, Advances in Production Engineering & Management, 11 (2016) 1, 38–48, doi:10.14743/apem2016.1.208 22 M. Hrelja, S. Klan~nik, T. Irgoli~, M. Pauli~, Z. Jurkovi~, J. Bali~, M. Brezo~nik, Particle swarm optimization approach for modelling a turning process, Advances in Production Engineering & Mana- gement, 9 (2014) 1, 21–30, doi:10.14743/apem2014.1.173 23 M. Hrelja, S. Klan~nik, J. Bali~, M. Brezo~nik, Modelling of a Turn- ing Process Using the Gravitational Search Algorithm, International Journal of Simulation Modelling, 13 (2014) 1, 30–41, doi:10.2507/ IJSIMM13(1)3.248 M. KOVA^I^ et al.: INCREASING THE TENSILE STRENGTH AND ELONGATION OF 16MnCrS5 STEEL ... 888 Materiali in tehnologije / Materials and technology 51 (2017) 6, 883–888 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS J. VANÌREK et al.: DURABILITY OF FRP/WOOD BONDS GLUED WITH EPOXY RESIN 889–895 DURABILITY OF FRP/WOOD BONDS GLUED WITH EPOXY RESIN OBSTOJNOST FRP/LESNIH SKLOPOV, LEPLJENIH Z EPOKSI SMOLO Jan Vanìrek, Milan [mak, Ivo Kusák, Petr Misák Brno University of Technology, Faculty of Civil Engineering, Veveøí 95, 602 00 Brno, Czech Republic vanerek.j@fce.vutbr.cz Prejem rokopisa – received: 2016-11-18; sprejem za objavo – accepted for publication: 2017-04-20 doi:10.17222/mit.2016.321 The paper describes the properties of FRP/wood bonds, made of different wood species commonly used in the timber industry within Central Europe (oak, spruce, pine and larch). An FRP fabric (glass-fiber-reinforced polymer – GFRP, carbon-fiber-reinforced polymer – CFRP) was applied as the reinforcement. The durability of all the reinforced FRP/wood assemblies was verified with short-term exposure tests following the tensile-shear-strength and wood-failure criteria for the shear area of a single-lap joint. To precisely define the effect of the mechanical interlocking mechanism of the bond, analyses of porosity and surface roughness, and SEM were performed. Keywords: durability, epoxy, FRP, wood adherend ^lanek opisuje lastnosti FRP/lesnih sklopov, narejenih iz razli~nih vrst lesa, ki se obi~ajno uporablja v lesni industriji v Srednji Evropi (hrast, jelka, bor, macesen). FRP-materiali (polimeri, oja~ani s steklenimi vlakni – angl. GFRP, polimeri, oja~ani z ogljikovimi vlakni, angl. CFRP) so bili uporabljeni za oja~anje. Obstojnost vseh FRP/lesnih sklopov so preverjali s kratkotrajnimi stri`nimi preizkusi. Kriterij je bil povr{ina stri`nega preloma na enojni prekrivni ploskvi med FRP in lesom. Da bi lahko natan~no definirali mehansko trdnost spoja, so izvedli analize poroznosti, povr{inske hrapavosti in SEM-analize. Klju~ne besede: trajnost, epoksi, FRP, lepljivost lesa 1 INTRODUCTION The main purpose of the FRP usage with timber in the construction industry is generally to improve the stiffness/strength of reinforced items without any influence on their service life or any environmental impact. From the perspective of the timber-rein- forcement process, the optimum dimensional stability during moisture changes in wood should be one of the most important criteria for such joints. Different studies were made for the FRP/wood bond durability using different types of adhesives. The optimum results of adhesion were achieved using formaldehyde-based adhesives;1–3 different, contrary results were found using epoxy resins.4,5 The results1 for the epoxy resins showed an inability to reach the requirement for a cohesive failure of 80 % (in wood) after an exposure to cyclic hygrothermal conditions. Moreover, positive results were found using priming treatments (hydroxymetyl resorcinol – HMR; resorcin-fenol – RF; hexamethylolmelamin metyl ether – MME) before the gluing process.2 The durability of FRP/wood joints using epoxy has not been completely established. Therefore, the final aim of this experiment was to establish the durability aspect of such joints. The focus was on the observation of the wood adhering parameters having a significant impact on the mecha- nical interlocking process to understand the impact of mechanical interlocking on the joint durability. 2 EXPERIMENTAL PART 2.1 Materials Carbon, glass or aramid are common reinforcement- fiber materials used for external polymer-composite systems known as fiber-reinforced polymers (FRPs). Reinforcing of wood with FRP lamellas/fabrics could be applied without any restriction regarding the wood species. To evaluate the influence of wood species, the experiment used glass-fiber-reinforced polymer (GFRP) and carbon-fiber-reinforced polymer (CFRP) fabrics applied to different wood adherents. Oak (Quercus robur), spruce (Picea abies), pine (Pinus sylvestris) and larch (Larix decidua) with a uniform thickness of 25 mm were chosen. The fabrics with carbon fibers (Tyfo SCH-41, FYFE Co.) with a width of 1.0 mm and the GFRP with glass fibre (Tyfo-SEH-51A, FYFE Co.) with a width of 1.3 mm were used. As the adhesive, the epoxy resin TYFO S (diglycidyl ether of bisphenol-A, DGEBA) with an amine hardener was used. Pre-treat- ment including penetration of the wood surface to ensure stabilization of wood cells (200 g/m2) was applied. Similarly, the FRP fabrics were saturated with epoxy resin (400 g/m2) on both sides. After 30 min of such Materiali in tehnologije / Materials and technology 51 (2017) 6, 889–895 889 MATERIALI IN TEHNOLOGIJE/MATERIALS AND TECHNOLOGY (1967–2017) – 50 LET/50 YEARS UDK 67.017:674-419.3:678.686 ISSN 1580-2949 Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 51(6)889(2017)