Study Of Rapid Prediction Method For Mortar Strength

In present China, the mortar used in architectural buildings is mainly prepared inconstruction site; therefore, the volatility of its property and quality is quiteconsiderable, which is dependent on the environment and construction conditions. Inorder to control the quality of mortar timely, one good way is to detect the property ofthe field-produced mortar as early as possible, which is significant for constructionenterprises to reduce their quality risk. In this paper, a systematic study isimplemented on rapid prediction method for mortar strength, especially for thestrength of cement mortar and mixed mortar. The content of the research is describedas follows:(a) An creative method known as “coagulant steam curing method” is raised topredict the mortar strength, referring to the early-stage prediction of concrete strength.Firstly, the coagulant CS is quantitatively added to the mass-produced mortar mixture.After being stirred evenly, the mixture is poured to get cubic specimens with size of40mm×40mm×50mm. Then, these specimens are carefully cured with hightemprature steam. According to scheduled time, mortar specimens are taken out to testtheir compressive strength values which are utilized to predict the strength of mortarthat is cured for28days under standard conditions. Herein, the sequences arepresented as "Technical procedures for rapid prediction of mortar strength".(b) The mix proportion design, specimen geometry design, accelerated curingexperiments and standard curing experiments are conducted on cement mortar as wellas mixed mortar with common strength grades (M5, M7.5and M10). The laboratoryexperiments totally consist of1080samples belonging to60categories.(c) Mathematical models, including unary linear model, binary linear model andternary linear models, are established between accelerated curing strength andstandard curing strength of mortar. Accordingly, corresponding strength equations areobtained through numerical regression analysis using MATLAB software. The unarylinear model, comprising one independent variable (i.e. accelerated curing strengthvalue), is convinient and time-saving for predicting the strength of commonly-usedmortar. While the binary linear model consists of two early-stage strength variables(e.g.3.5h and4.5h strength values) but the prediction is much more accurate. Besidesthe3.5h and4.5h strength values, the ternary linear model also expropriates mix ratioas one variable in mathematical equation; thus, the model could predict the strength of mortar with different strength index or different type.(d) The fundamental principles of Artificial Nuero Netwoks (ANN) areintroduced in this paper, and then the application of ANN on mortar strengthprediction is further discussed. On the platform of MATLAB, the BP Nureo Networksmodel is established to predict the strength of cement mortar and mixed mortar.Meanwhile, the capacity of this predicting model is tested.(e) With a longer curing time (3.5h and4.5h), the strength values show a goodcorrelation and smaller discreteness. Therefore, during construction, we are supposedto use the strength value with a longer curing time to predict the28d strength toimprove the accuracy.The strength development laws of cement mortar and mixedmortar wtih curing time under coagulant steam condition are explored. In accordancewith the experimental results, the strength as well as the strength ratio of mortar showincreasing trend with curing time under high-temprature steam condition. If thedosage of cement increases, the absolute amount of cement that participated in thesetting and hardening reaction would increase, while the relative amount will increaseless.