Effect Of Technological Parameters Of Production Of Tfp Preforms Of Gas Turbine Engine Blades On Their Physical And Mechanical Properties

The article discusses the features of the destruction of carbon fiber reinforced plastics on the basis of tfp-preforms produced using various stitching mode parameters, and provides for studies of their physical and mechanical properties and structure. The effect of the carbon fiber laying density and the piercing step of the aramid yarn on the elastic-strength properties of composites is shown. The optimal technological parameters for creating tfp-preforms of gas turbine engine blades were chosen.


Introduction
Carbon fiber reinforced plastics (CFRP), the preforms of which are obtained by the tailored fiber placement (TFP) technology [1,2], represent a promising class of materials with a complex of unique properties [3,4]. Despite the fact that the TFP technology for the production of composites is largely unknown in Russia, the world experience of its application [5][6][7] allows to call it the most promising technology among other methods for creating 3D preforms.Preform stitching is carried out on industrial numerically controlled sewing machines. The machine automatically moves the carrier material in accordance with a software template specifically developed for produced TFP preforms.However, in the manufacture of such materials, due to the penetration of the needle of the sewing machine, defects are formed in the reinforcing structure, which can significantly reduce the elastic-strength characteristics of reinforced plastics. Since defects in patch composites can become critical and lead to premature destruction of the material, studies of the physical, mechanical and technological properties of patch reinforced plastics and their destruction mechanisms are very relevant. When creating preforms by automated laying of the carbon fibers on a water-soluble substrate, the physical and mechanical properties of finished products are affected not only by the reinforcement scheme and the properties of the materials used (carbon and aramid fibers, substrates). To achieve optimal elastic-strength properties, the technological preform stitching parameters (roving laying density ρand the piercing step Δ) that are currently poorly studied should be also taken into account. This is particularly important for high strength parts, for example, for gas turbine engine blades (GTE blades).
It is necessary to select the optimal parameters of TFPprocess so that the maximum transverse characteristics are achieved without significantly reducing the strength along the direction of reinforcement, which is an important factor for GTE blades.
The objective of this study, which is continuation of the authors' research in [8], is to determine how the roving laying density and the piercing step of carbon fiber stitching affect the transversal properties of model flat samples and GTE blades samples, as well as to consider the structure and features of fracture of reinforced plastics on the basis of the tfp-preforms.

Materials and methods
Stitched carbon fiber reinforced plasticswere made on the basis of HTS45 12K carbon fiber (TohoTenax®, Japan). The piercing thread was an aramid thread Rusar-S (LLC "Thermotex", Khotkovo, Moscow region, Russia). Preform stitching was carried out on the JCW 0100-500 automatic numerically controlled sewing machine (ZSK Stickmaschinen GmbH, Germany). All manufactured preforms were reinforced in the directions [0°; 90°], the laying density ρ varied from 47 to 70 CU, the piercingstep Δ -from 4 to 10 mm. Control samples of carbon fiber were obtained by winding carbon fiber HTS 45 12K (TohoTenax ® , Japan) on a plate in the direction [0°;90°] and pressed.
All preforms were impregnated by VaRTM at a room temperature of 23°C and a pressure of -1 atm for 30 minutes. After impregnation the preforms were cured for 4 hours at 110 °C.
The resulting reinforced plates were cut into samples of the required sizes for testing for short-beam bending (at quasi-static and dynamic loading speeds), tension and compression using the universal Zwick Roell Z100 testing machine (Zwick GmbH & Co. KG, Germany). The structure was studied using a Bruker SkyScan 1172 microtomograph and a PhenomPro X scanning electron microscope. The three-point bending test [9] under quasi-static loading was carried out on flat samples 8 mm in width, 44 mm in length and 3÷4.4 mm in thickness. The loading rate was 10 mm/min. To exclude the effect of shear stress during bending, the distance between the supports lwas chosen so that the ratio l to the sample thickness of more than 20 was achieved. The bending strength σs(MPa) was determined according to the formula: (1) whereFmax is the load corresponding to the destruction of the sample, b is the sample width, h is the sample thickness, and l is the distance between the supports. The total fracture energy Afswas calculated according to the formula: where the integral expression is the area under the load-deflection curve from the initial deflection l0 to the deflection li, b is the sample width, and h is the sample thickness. The three-point bending test [10] under of low-velocity impact conditions was carried out on flat samples 8 mm in width, 44 mm in length and 3÷4.4 mm in thickness. The movement rate of the impactor was 4 m/s (impact energy of 13 J). The bending strength σd(MPa) was determined according to the formula (1). The total fracture energy Afdwas calculated according to the formula: wherev is the velocity of the impactor at the time of contact with the sample, m is the weight of the impactor, the integral expression is the area under the load-time curve from the initial moment of loading t0 to ti. The four-point bending test [9] was carried out on flat samples 14 mm in width, 76 mm in length and 3÷4.4 mm in thickness. The loading rate was 10 mm/min. The bending strength σb(MPa) was determined according to the formula (1). The elasticity modulus Ebwas calculated according to the formula: = 0,21 ∆ 3 whereΔω is the difference in the deflections ω` and ω``, ΔF is the increment of load in the elastic section of the load corresponding to the change in the deflection Δω. The torsion test [11] was carried out on flat samples 80 mm in width, 80 mm in length and 3÷4.4 mm in thickness. The loading rate was 1 mm/min. The shear modulus of elasticity G was determined according to the formula: where F is the load applied to the corner of the sample, h is the sample thickness, l is the distance between the supports, and Δl is the deflection under the point of force application. Using the loading diagram by the least squares method, the coefficient k was determined in the linear equation y=kx, where y=F, and k=Δl. With regard to the found coefficient k, the shear modulus G was calculated according to the formula: (6) The tensile strength was calculated using the formula: The compressive strength was calculated using the formula: The statistical processing of the measurement results was performed for each type of research by calculating the mean value and the standard deviation. Figure 1 shows typical photos of destroyed carbon fiber plastics during the three-point bend of a short beam. It can be seen that the destruction of the sewn samples (Figure 1a-1d) is localized in the central area. This type of cracking is typical for carbon fiber reinforced plastics that break when bent from normal stresses. However, in the central part of the samples longitudinal cracks are also observed, which are formed during destruction from shear stresses. In this case, the movement of cracks from shear stresses ends near the aramid fibers, which is explained by their greater resistance to crack growth. This type of destruction was typical for the entire range of technological parameters of the stripe. In the case of samples obtained by pressing, typical failures were observeddelaminations in the central area from shear stresses (Figure 1e). In typical loading diagrams (Figures 2 and 3) forthree-point bending a fairly high level of peaks corresponding to secondary failure was observed. It can be seen that for pressed materials the crack spread in spurts, as evidenced by high secondary load peaks. For stitched carbon fiber plastics the load fluctuation with the growth of deformation was significantly less. This may indicate that the crack in such materials spreads smoothly, without significant jumps. This mechanism of destruction is associated with the presence of aramid fiber in the structure of carbon fiber reinforced plastics. In this case, the aramid thread slows down the movement of cracks and their further development. It should be noted that the type of loading diagrams practically remained the same when changing the piercingstep or the density of carbon thread stitching. For the four-point bend a mixed failure mechanism of the stitched samples was also observed. Figure 4 shows that the samples were destroyed from both normal and shear stresses, as evidenced by cracks that spread along and across the carbon fiber reinforced plastics. In this case, the samples are destroyed by normal stresses to a greater extent, while the development of cracks from shear stresses is inhibited by the stitching of aramid fibers. The density and piercingstep also have little effect on the type of destruction and are random. For the pressed carbon fiber plastics a typical type of failure was observed for such materialsa crack in the central area from normal stresses.    (Table 1). No clearly expressed dependences of the E and G moduli on the stitching step or density could be obtained. However, a trend toward increasing the elasticity and shear moduli can be seen at a piercing step of 7 to 10 mm for any stitching density.  Figure 9 shows the results of measuring the tensile and compressive strength. It can be seen that the strength of the stitched carbon plastics in both cases was lower than the pressed ones, and the tensile strength was almost 2 times. At the same time, compression is less sensitive to defects. Table 2 presents the comparative properties of pressed and stitched CFRP obtained within the optimal range of the technological parameters. It is apparent that the manufacture of products using the TFP technology allows to approach the properties of highly oriented pressed composites with high strength characteristics.  Figures 10 and 11 show the reinforcing structures of stitched plastics obtained on microtomograph and scanning electron microscope. It can be seen that all technological modes of obtaining carbon fiber plastics are characterized by defects in the place of penetration of the material by the needle of the sewing machine. When the needle is pierced, some of the carbon nanofibers are destroyed, while the other part is bent. The formation of a loop of aramid thread contributes to the formation of pores. The pore distribution in stitchedcarbon fiber plastics requires a separate study. Products for which high physical and mechanical properties determine the resource and reliability of their operation include gas turbine engine blades. To confirm the suitability of the developed technology, compressor blade preforms were produced for the centrifugal pump impeller (CPI). Table 3 shows the test results for the produced GTE compressor blades the preforms of which were obtained by manual laying of the carbon fabric on the basis of the T700 and IMS 65 carbon fibers, as well as by directional stitching of the IMS 65 carbon roving with parameters of ρ=57 CU and Δ=7 mm. The bending strength for the stitched blades was 25-30 % lower than for non-stitched carbon fiber reinforced plastics. Nevertheless, taking into account the fact that it is extremely difficult to implement an oriented fibrous structure in stitched preforms, the product maintaining the sufficient bending strength and elasticity of the samples significantly (up to 30 %) won in the shear strength.

Results and discussion
An increase in the ultimate shear strength even by 10 % in the reinforcement plane will allow to preserve to a greater extent the integrity of the finished product, preventing it from disintegrating into fragments even under operating modes close to destruction. Such a parameter is very important for GTE elements as soon as fragments formed even in a small amount often have a very high kinetic energy and are able to damage the rest engine parts.

Conclusion
The conductedstudies of the effect of the technological parameters on the physical and mechanical properties of CFRP suggest that the properties of stitched materials are consistent with the properties of pressed CFRP with a highly oriented structure the bending strength of which differs by not more than 25 %. Despite the fact that sewn carbon fiber plastics have strength characteristics 20 -50% lower than pressed ones, it should be noted that under dynamic loading, sewn plastics have a higher energy of destruction than pressed ones. The increased energy of destruction is associated with the presence of aramid fibers in the reinforcing structure, which have a greater dissipative capacity compared to carbon fibers. The most energy consuming structure and increased elastic-strength characteristics of stitched CFRPare realized atthe stitching step of 7-10 mm and roving laying density of 50-60 CU, which were selected as optimal for the production of GTE blades. One can hope that the crack growth in such materials and GTE blades in particularwill not be so disastrous even under impact loads in contrast to pressed carbon fiber reinforced plastics.
In addition to the achievement of a combination of physical and mechanical properties that are optimal for operation, the directional fiber stitching method possesses such advantages as the possibility to produce series of products with identical geometric, weight and strength characteristics, the absence of a human factor, and the efficient and non-waste production, which makes it promising for application in the aerospace industry.