Mechanical properties of recycled micropowder engineered cementitious composites in cold region based on BP neural network
The mechanical properties of recycled micro powder engineered cementitious composites(ECCs)un-der various low-temperature conditions are investigated.The control variable method explores the ECC's com-pressive and flexural strength concerning two low-temperature conditions:various freeze-thaw cycles and con-stant negative temperatures,with different types of recycled micro powders and replacement rates as variables.The effects of recycled micro powder type,replacement rate,number of freeze-thaw cycles,and negative tem-perature on the mechanical properties of ECC were analyzed.Moreover,two prediction models were presented to calculate their compressive strengths based on 3-6-1 and 3-3-1 BP neural network structures for two conditions of freeze-thaw cycle and constant negative temperature.Results show that the mechanical strengths of recycled concrete micropowder ECC are higher than those of recycled brick micro powder ECC under the same freeze-thaw conditions.With the increase in the replacement rate of recycled micro powder,a tiny decline and a sharp decline were observed,as well as a 20%loss of mechanical performance for 150 freeze-thaw cycles.However,the mechanical properties of the recycled micro powder ECC showed the opposite trend after experiencing con-stant low-temperature insulation.The recycled micro powder ECC mechanical strength showed an upward trend of 22%,from room temperature to-40℃constant low-temperature state.The two BP neural network prediction models were established under low-temperature conditions with an average relative error of 1.43%and 1.28%,respectively.The maximum freeze-thaw cycles of recycled micro powder ECC with different mix ratios were predicted based on the mass loss rate and relative dynamic modulus.
NETL
NSTL
万方数据
