Microstructure Characterization Of Fresh Cement Paste And Its Correlation With Rheology

Cement is widely used in the concrete industry.The rheology of cement paste will directly affect the early workability of concrete.The essence of the rheology of cement paste is the change of the microstructure,and the addition of superplasticizer will significantly improve its early rheological properties.In essence,the superplasticizer adsorption can change the interaction between particles to optimize the microstructure,so the rheology of cement paste is closely related to its microstructure.This study takes fresh cement paste as the research object,and aims to reveal the relationship between the rheology of fresh cement paste and its microstructure.The thesis systematically studied the rheological properties and microstructure of fresh cement paste under the action of different types and different amounts of superplasticizers,and introduced the theories of spectrum analysis and fractal geometry into this field.Several reasonable,sensitive and stable parameters have been extracted to comprehensively and quantitatively characterize the microstructure of the paste.At the same time,based on the macro and micro parameters of the cement paste,a neural network prediction model of the rheological parameters of the fresh cement paste was established.Studies have shown that the superplasticizer changes the microstructure of cement paste through adsorption and flocculation,which affects the rheology of the paste;as an active reaction system,the continuous progress of cement hydration and temperature changes will also affect the microstructure thus causes its rheology to change constantly.The main results and conclusions are as follows:1)Through the rheological test,the rheological change rule of cement paste under the influence of the type,dosage,temperature and time of the water reducing agent is clarified;under the joint action of the water reducing agent adsorption and cement hydration,the rheological parameter value varies with the the time is rising.With the increase of the dosage,more and more water-reducing agents have a smaller and smaller change with time.The viscosity value under the action of some superplasticizer does not even change with time.2)Corresponding characteristic factors-initial value,initial time variability and later time variability are extracted from the obtained rheological parameters.These characterizing parameters can clearly reflect the changing rules of the rheological parameters,thus deepening the understanding of the rheological changes.It also provides a new processing method for data analysis of rheological properties.Among them,the overall initial value increases with the increase of the amount of superplasticizer;the overall initial time-varying rate decreases with the increase of the amount of superplasticizer;the overall time-varying rate with the increase of the amount of superplasticizer later decrease first and then stabilize.3)Using image analysis and fractal theory to quantitatively describe the microstructure characteristics of the paste,and extract the corresponding microscopic characterization parameters.The average particle size f_d directly obtained by the Morphologi G3 microscope characterizes the size of the cement particles;the peak of the spectrum platform f_p obtained by spectrum analysis,the width of the spectrum platform f_w,and the median of the spectrum platform f_m all include the cement particles in the microscopic images to varying degrees uniformity and periodicity information;the fractal dimension ff obtained by the fractal dimension theory calculation characterizes the self-similarity of the microstructure;the shape factor f_s obtained by the graphical approximation formula is derived to comprehensively characterize the shape information of the cement particles.4)The gray correlation analysis of the microstructure characterization parameters,plastic viscosity and yield stress was carried out,and the correlation degree of each microstructure parameter with plastic viscosity and yield stress was obtained.Based on this,the microstructure parameters were extracted as the input of the BP neural network layer,the neural network prediction model of plastic viscosity and yield stress is established.