Experimental research on mechanical properties of desert sand steel-PVA fiber engineered cementitious composites

An orthogonal experimental design method involving five-factor and four-level is adopted for the mix design of Desert Sand Steel-PVA fiber ECC. The effect of each level on Mechanical properties of ECC and the difference of Mechanical properties between each level is analyzed. The influence of different experimental factors is discussed, which includes water-binder ratio, fly ash substitution rate, desert sand substitution rate, proportion of PVA fiber and proportion of steel fiber. The experimental results indicate that water-binder ratio and fly ash substitution rate are the most principal and significant influencing factors on the compressive strength of ECC, regardless of age. Steel fiber is conducive to development of splitting tensile strength; PVA fiber is conducive to the development of flexural strength. High strength ECC can be prepared when the desert sand substitution rate is high. As the raw material of ECC, river sand can be 90% replaced by desert sand.


Introduction
Engineered Cementitious Composites (hereinafter referred to as ECC) was proposed in the last late century ECC is micromechanically designed composites with fiber as reinforcement material, cement as main base material. It features outstanding energy consuming capacity, strain hardening and crack steady state development. ECC can effectively improve the seismic resistance of the structure [1], extend service life of the material [2].
Research on steel-PVA cementitious composites show that addition of fly ash, increase of fiber proportion help multiple cracking [3.4]. PVA-ECC has outstanding property of peak delay [5]. The damage crack of PVA-ECC is less wide, which can be controlled within 100μm [2,6], Harbin Institute of Technology researched high fly ash proportion ECC, its 28d limit tensile strain is all above 3%, exhibiting outstanding strain-hardening property [7]. From the angle of sustainability of the environment, ECC consumes less resource, emits less pollution [8]. Adoption of ECC can increase compressive strength and tensile strength of the concrete test piece [9]. Experiments show that compressive strength of 4-day aged ECC can meet the strength required by deck slab design, conducive to speeding up construction [10]. Using micromechanical design can achieve ideal performance of ECC [11]. Fly ash, water-binder ratio, fiber disperse condition all can influence tensile property of ECC [12,13]. When steel fiber proportion is 2.0%, difficult dispersal and serious caking of steel fiber results strength of ECC lower than that of the test piece with steel fiber proportion 1,5% [14].
Large amount of non-renewable building material is consumed, so using a substitute for building material is of significance for sustainable development of the industry. Sand for building has limited reserve and cannot be regenerated within short term worldwide, while every continent has deserts of different size, and desert sand can partly substitute for river sand used for concrete, whose smaller grain size makes concrete inside more even, more compact. Currently, the thermal power is still the main electric source globally, nowhere can place a large amount of waste fly ash discharged from thermal power plants. Fly ash can substitute for cement to a certain extent, and improve performance of concrete, avoid waste pollution. Desert sand and fly ash are used in concrete as substitute material, saving resources and lowering cost. Desert sand steel-PVA fiber ECC prepared in this paper uses desert sand to substitute for part of river sand, fly ash to substitute for part of cement.

Experimental material
Saima branded p.o.42.5 ordinary portland cement of Ningxia, China, its specific surface area is 339m 2 /kg, the chemical composition is shown in Table 1, the properties is shown in Table 2; for class I fly ash of Ningxia Lingwu Thermal Power Plant, water content is 0.4%, water demand ratio is 90%, fineness is 8.4%, loss on ignition is 3%, the chemical composition is shown in Table 1; the desert sand is from Tengger Desert, with average grain size 0.23mm, fineness modulus is 0.7; for sieved river sand, max grain size is 1.18mm, fineness modulus is 2.1; PVA fiber is produced by Japan KURARAY, steel fiber selects copper plated micro wire steel fiber produced by Hengshui Fangde Silk Screen Products Factory, performance of fiber is shown in Table 3. Admixture is powdered high efficiency polycarboxylic acid water reducer, with water reducing rate 25%~30%.

Proportion of experiment
The paper uses 5-factor 4-level (L 16 4 5 ) orthogonal experiment, factors and levels is shown in Table 4.
Mix design of orthogonal experiment is shown in Table 5.
solid. Strip mould after 24h, being cured to 7d, 28d. After curing, dry the test piece and conduct mechanical property experiment. Measure failure load on universal testing machine, calculating compressive strength, splitting tensile strength, flexural strength according to corresponding equation respectively.
The compressive strength was measured by the 70.7mm ´ 70.7mm ´ 70.7mm cube specimens, the tensile strength was measured by the 100mm ´ 100mm ´ 100mm cube specimens, and the flexural strength was measured by 40mm ´ 40mm ´ 160mm prism specimens.

Experimental result and analysis
Experimental results of 7d, 28d compressive strength and splitting tensile strength and 28d flexural strength are shown in Table 6. The purpose of this experiment is to study the effect of individual factors on the properties of ECC, and the combination of the interaction between the factors is complex, so the interaction between factors is not discussed in this paper.

Intuitive analysis
It is known from No. 3 and No.7 28d compressive strength in Table 6 that: When water-binder ratio is the lowest (0.19), fly ash substitution rate is 45%, when using high desert sand substitute (60%), we can still prepare high strength ECC with compressive strength above 75Mpa (No. 3); when waterbinder ratio is low (0.24), even fly ash substitution rate is up to 45%, using extremely high desert sand substitution rate (90%), the compressive strength is still up to 60 Mpa (No. 7).

Range analysis
Range analysis is shown in Table 7. The influence trends of each factor in the strength is shown in Figure 1.
It is known from Table 7 and Figure 1 that: (1) As water-binder ratio increases, compressive strength, splitting tensile strength and flexural strength show decreasing trend; increase of the fly ash substitution rate will result in a decrease of 3 properties. As desert sand substitution rate increases, compressive strength and splitting tensile strength both increase first and decrease slowly then. Flexural strength shows the trend of slow decrease. When the desert sand substitution rate is 30%, compressive strength and splitting tensile strength reach the highest; as Proportion of PVA fiber increases, compressive strength and splitting tensile strength increases first and decreases then, when the proportion is 0.8%, compressive strength and splitting tensile strength reach the highest; flexural strength increases with progressive increase of proportion of PVA fiber. As a proportion of steel fiber increases, compressive strength increases first and de- (2) For 7d compressive strength, ranking of every influencing factors is: fly ash substitution rate (B)> water-binder ratio (A)> Proportion of PVA fiber(D) > desert sand substitute (C) > Proportion of steel fiber(E), the better combination of condition is A2B1C2D3E2. For 28d compressive strength, ranking of every influencing factors is: water-binder ratio (A) > fly ash substitution rate (B) >desert sand substitute (C) > Proportion of PVA fiber(D) >Proportion of steel fiber(E), the better combination of condition is A1B1C2D3E2. For 7d splitting tensile strength, ranking of every influencing factors is: waterbinder ratio (A) >Proportion of steel fiber(E) > Proportion of PVA fiber(D) > fly ash substitution rate (B) >desert sand substitute (C), the better combination of condition is A1B1C2D3E4.
For 28d splitting tensile strength, ranking of every influencing factors is: Proportion of steel fiber(E) >fly ash substitution rate (B) >waterbinder ratio (A)>Proportion of PVA fiber(D) >desert sand substitute (C), the better combination of condition is A1B1C2D3E4. For 28d flexural strength, ranking of every influencing factors is: Proportion of PVA fiber(D) > Proportion of steel fiber(E) >fly ash substitution rate (B)>water-binder ratio (A) >desert sand substitute (C), the better combination of condition is A1B1C1D4E4.
(3) The influence of water-binder ratio and fly ash substitution rate of compressive strength is especially significant; 7d splitting tensile strength is influenced by water-binder ratio most; 28d splitting tensile strength is influenced by Proportion of steel fiber most; (4) According to integrated balance method, better combination of conditions is determined as A1B1C2D3E4.

Variance analysis
Variance analysis is shown in Table 8. Taking into account the four-level and five-factor orthogonal test, in the variance analysis, the factor whose mean square is the smallest is considered as error to calculate the value of F. It is known from Table 8 that: -According to mean square value, rank significance of influence of every factor on mechanical property from large to small, for splitting tensile strength, flexural strength, 7d compressive strength, the ranking is in accordance with range analysis; for 28d compressive strength, ranking of influence of every factors is: waterbinder ratio (A) >fly ash substitution rate (B) >Proportion of PVA fiber(D) >desert sand substitute (C) >Proportion of steel fiber(E), slightly different from range analysis.

Compression failure
The specimen without fiber often breaks unexpectedly, with the larger piece falling off as shown in Figure 2 (a). The specimen with only steel fiber has small piece falling off, with relatively complete contour, the surface shows several cracks with certain width, as shown in  Figure 2(b). The specimen with steel-PVA fiber has many micro cracks in the surface, with almost nothing falling off in the surface, keeping good integrity, as shown in Figure 2(c), with cracks marked with marker pen.

Splitting failure
The specimen without fiber will suddenly split with a "Bang" under max load, totally breaking, as shown in Figure 3(a). The specimen with only steel fiber will form several wider main cracks, and the specimen is still linked together, as shown in Figure 3(b). The specimen with steel-PVA fiber forms 1 or two thinner main cracks and several micro cracks, failure mark in the surface is not apparent, as shown in Figure 3(c).

Flexural failure
The specimen without fiber gives out crisp sound of breaking when failing, the test piece is bent to two halves totally, as shown in Figure 4(a). The test specimen with only steel fiber forms a wider crack in tensile area, the compressive area is still linked together, as shown in Figure 4(b). The specimen with steel-PVA fiber breaks to smaller cracks in compressive area, with relatively integrated form, as shown in Figure 4(c).

Cause analysis
Test pieces with different type of fiber have different failure modes(There is no significant difference between specimens with only PVA fiber and specimens with steel-PVA fiber, so the failure mode of a specimen with only PVA fiber is not presented in the text). Two kinds of fiber used in the experiment are different greatly in size of the diameter, playing bridging role in different scale. PVA fibers mainly controls micro crack at an early stage of bearing load, steel fiber mainly controls macro crack. The test piece without fiber will generate a fragile burst failure; steel fiber ECC typically produces several wider main cracks when it fails; steel-PVA fiber ECC will produce many micro cracks when it fails, the form is damaged less, failure is slow, showing better ductility.

Conclusion
Using inexhaustible desert sand and industrial waste fly ash, through orthogonal experiment, this paper mainly researches influence trend of 5 factors, including water-binder ratio, fly ash substitution rate, desert sand substitution rate, proportion of PVA fiber, and proportion of steel fiber on ECC strengths, in order to find main factor influencing strength. According to failure mode of test pieces, the role   played in desert sand ECC by different fiber is analyzed. And it is concluded that: -When water-binder ratio is low, even if adopting high fly ash substitute and adding large amount of desert sand, high strength ECC can still be prepared, which is of significance for desert sand ECC at key location of high rise anti-seismic structure.
-According to integrated balance method, finalized better condition of factor combination is A1B1C2D3E4, namely water-binder ratio is 0.19, fly ash substitution rate is 15%, the desert sand substitution rate is 30%, proportion of PVA fiber is 0.8%, proportion of steel fiber is 1.2%.
-Water-binder ratio and fly ash substitution rate influence compressive strength highly significant (a = 0.01), but low water-binder ratio and fly ash substitution rate result in even larger compressive strength; proportion of steel fibers significantly (a = 0.05) influences 28d splitting tensile strength, and steel fiber is conducive to development of splitting tensile strength; PVA fiber is conducive to the development of flexural strength.
-Desert sand substitution rate does not significantly influence every mechanical properties, but appropriate desert sand substitution rate can improve ECC property to a certain degree. In addition, max desert sand substitute in the experiment has been up to 90%, and the experimental group adopting 90% desert sand substitution rate does not decrease obviously in compressive strength, splitting tensile strength compared with the experimental group without desert sand, but flexural strength decreases to some extent. So it is inferred that as raw material of ECC, desert sand can further totally substitute for river sand.
-PVA fiber enhances ductility of desert sand ECC, making the failure process of test piece failure intend to slow, conducive to multicrack development of failure.