Cement strength prediction analysis method based on deep fusion of multi-source data

CN122243911APending Publication Date: 2026-06-19DONGPING ZHONGLIAN MEIJING CEMENT CO LTD

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Applications(China)
Current Assignee / Owner: DONGPING ZHONGLIAN MEIJING CEMENT CO LTD
Filing Date: 2026-03-16
Publication Date: 2026-06-19

AI Technical Summary

Technical Problem

Existing cement strength prediction models rely on macroscopic physicochemical indicators, which cannot reveal the relationship between microstructure and defect structure. This leads to a decrease in prediction accuracy when faced with strength anomalies caused by micro-defects, and the models cannot be traced back to the production process, thus weakening their guiding effect on the production process.

Method used

By acquiring microstructural images of hardened cement bodies, potential deterioration areas are identified, porosity and three-dimensional connectivity are analyzed, and spatiotemporal correlation is performed by combining full-process time-series data to identify key process fluctuation points and generate process parameter fine-tuning instructions, thus achieving closed-loop decision-making from microstructural diagnosis to process root cause tracing.

Benefits of technology

It provides a basis for reflecting the internal microstructure of cement, reveals the fundamental process causes of strength loss, and realizes closed-loop control from microstructure diagnosis to process parameter optimization, thereby improving the accuracy of prediction and the pertinence of production adjustments.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122243911A_ABST

Patent Text Reader

Abstract

This invention discloses a cement strength prediction and analysis method based on deep fusion of multi-source data, relating to the field of cement strength prediction and analysis. The method includes: acquiring a three-dimensional microstructure image of the interior of a hardened cement sample, extracting porosity based on the image, identifying potential deterioration areas, analyzing pore size distribution and three-dimensional connectivity, and calculating a pore deterioration feature vector; performing spatiotemporal correlation and pattern matching between the pore deterioration feature vector and the full-process time-series data of the corresponding production period to identify key process fluctuation points leading to specific microstructural defects; making predictions based on the correlation between the pore deterioration feature vector and historical strength loss data, and generating process parameter fine-tuning instructions based on strength deviation values and key process parameter types. This achieves a closed loop from intelligent diagnosis of microstructural defects to precise tracing of production process fluctuations, and finally, control of process parameters.

Need to check novelty before this filing date? Find Prior Art