Blast-Induced Ground Vibration (PPV) Prediction Using USBM Scaling Law and Machine Learning

Every quarry blast sends ground vibration to nearby houses, and the peak particle velocity (PPV) at the nearest structure is limited by regulators. Blast engineers predict PPV from the distance and the charge with empirical scaling laws and choose a safe charge. This project fits the six classical laws (USBM square-root scaled distance, Ambraseys-Hendron, Langefors-Kihlstrom, Indian Standard, CMRI and a free power law) to 200 real blast records of ten Nigerian quarries and compares them with machine learning (Gaussian process, Random Forest, XGBoost with and without monotone constraints, a neural network, a USBM + Gaussian-process hybrid) and PySR symbolic regression, under honest splits that hold whole quarries out. Every model is checked for physical consistency, the safe charge is given with a 95 % prediction bound, and a web application supports blast design, blast plans with approval and measurement, quarry laws re-fitted from monitoring and limit exceedances on five databases. Weka, PySR and GNU Octave re-check the results; a complete MATLAB version is included.

Category: Mining Engineering

Mining Engineering Project MN04 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.