A new application of radioactive particle tracking using MCNPX code and artificial neural network

Area of Science:

• Nuclear Engineering
• Industrial Process Monitoring
• Materials Science

Background:

• Stirrers and mixers are crucial in chemical, food, pharmaceutical, and concrete industries.
• Ensuring proper mixing and homogeneity is vital for product quality and competitiveness.
• Current diagnostic tools may not adequately assess mixture uniformity in real-time.

Purpose of the Study:

• To develop and validate a non-invasive diagnostic tool for industrial mixers.
• To predict the instantaneous position of a radioactive particle for monitoring concrete mixtures.
• To assess the feasibility of radioactive particle tracking for evaluating industrial mixing processes.

Main Methods:

• Utilized radioactive particle tracking (RPT) principles.
• Employed Monte Carlo simulations (MCNPX code) for modeling.
• Implemented artificial neural networks (ANNs) for position prediction.
• Simulated a detection system with eight NaI(Tl) detectors and a 137Cs source.

Main Results:

• Achieved a high correlation coefficient of 0.99 for all coordinates in both air-filled and concrete-filled tanks.
• Demonstrated the capability to accurately predict the radioactive particle's position.
• Validated the MCNPX code's effectiveness in modeling the detection system.

Conclusions:

• The proposed radioactive particle tracking methodology shows significant promise as a diagnostic tool for industrial mixers.
• This technique can effectively monitor mixture homogeneity and ensure product quality.
• Further development could lead to widespread industrial application for process optimization.