Designed strength identification of concrete by ultrasonic signal processing based on artificial intelligence...

Area of Science:

• Civil Engineering
• Materials Science
• Artificial Intelligence

Background:

• Accurate identification of concrete strength is crucial for structural integrity and safety.
• Traditional methods for assessing concrete strength can be destructive and time-consuming.
• Non-destructive testing methods are sought for efficient and reliable concrete evaluation.

Purpose of the Study:

• To develop and validate a pattern recognition method for identifying designed concrete strength.
• To utilize artificial intelligence techniques, specifically evidence accumulation, for concrete strength classification.
• To explore the effectiveness of ultrasonic signal features in concrete pattern identification.

Main Methods:

• Extraction of multiple feature parameters (variance, zero-crossing, mean frequency, AR model coefficients, linear cepstrum coefficients) from ultrasonic signals.
• Application of an evidence accumulation procedure for pattern recognition using measured distances.
• Introduction of a fuzzy mapping function to adapt distance measurements for the evidence accumulation method.
• Testing on concrete specimens with designed strengths ranging from 180 to 400 kg/cm2.

Main Results:

• The proposed pattern recognition method demonstrated feasibility in identifying concrete patterns.
• Ultrasonic signal features effectively contributed to the classification of concrete strengths.
• The evidence accumulation approach, enhanced by fuzzy mapping, proved suitable for this application.

Conclusions:

• The developed AI-based pattern recognition method offers a viable non-destructive approach for concrete strength identification.
• This technique can potentially improve the efficiency and accuracy of concrete quality control in construction.
• Further research can explore broader applications and refinements of this intelligent system for material assessment.