Preparation And Mechanical Properties Of Ultra-high-performance Lightweight Concrete

Ultra-high-performance concrete(UHPC)is an advanced cementitious material with a water-binder ratio of 0.15～0.25,and it has good workability,mechanical properties and durability.However,the self-weight of UHPC is rather high.Currently,ultra-high-performance lightweight concrete(UHPLC)characterized by both ultra-high strength and light weight,is one of the research hotspots in civil engineering materials.In this work,UHPLC mixture was designed through the Modified Andreasen and Andersen particle packing model.Shale ceramsite with different particle sizes(0～3mm,5～8mm,8～12mm)was used to prepare different kinds of UHPLC,and the workability,mechanical properties and microstructure of UHPLC were explored,and then the optimal particle size of shale ceramsite was obtained.Based on this,the feasibility of producing UHPLC by using air entrainment agent(AEA)was studied,and the effects of the types and contents of AEA on the workability,mechanical properties and microstructure of UHPLC were studied,which suggested that the apparent density of UHPLC was reduced but its mechanical properties were also deteriotated,due to the addition of AEA.Therefore,the graphene oxide(GO)was used to improve the mechanical properties of UHPLC,and the effects of GO on the mechanical properties and microstructure of UHPLC were investigated.Finally,the back propagation neuronal network model(BP-ANN)was used to predict the compressive strength of UHPLC.The main conclusions of this study were shown as follows:(1)When the particle size of shale ceramsite was 0～3mm,the prepared UHPLC has higher mechanical strengths and lower porosity.When the particle size of shale ceramsite was 5～8mm or 8～12mm,the apparent density of UHPLC was decreased,but its mechanical properties were also reduced,and its porosity was increased.(2)With the increase of AEA content,the apparent density and mechanical properties of UHPLC were decreased continually,but the prepared UHPLC has a higher specific strength,when the content of AEA was 0.01%.Weighing the requirement of light weight and high strength,the optimal content of AEA in UHPLC was 0.01%.(3)Due to the incorporation of GO,the flexural strength,compressive strength and elastic modulus of UHPLC were increased by 4.4%～21.9%,6.3%～22.4%and 6.1%～12.9%,respectively.The microstructure of UHPLC was improved,and the porosity of UHPLC were reduced,because of the addition of GO.In addition,the generation of high-density C-S-H and ultra-high-density C-S-H in UHPLC cement paste were also promoted due to the addition of GO,and thus the elastic modulus of UHPLC could be improved.Considering the influence of GO on the mechanical properties and microstructure of UHPLC,the optimal content of GO in UHPLC was 0.06%.(4)When the hidden layers of the BP-ANN model were 13 neurons,most of the errors between the predicted value and the experimental value were within 10%.The correlation coefficient R~2 of fiting curve between the predicted value and experimental value of UHPLC was0.96599,which indicated that this BP-ANN neural network model could be used to predict the compressive strength of UHPLC accurately.