Improved distance discriminant analysis model for slope blasting stability discrimination

Aiming at the problem that there are many uncertain factors in the evaluation of mine slope stability by blasting vibration, the distance discriminant analysis theory was adopted and the weighted value of the model was improved to distinguish the importance of different parameters. The distance discriminant analysis model (DDA model) and the improved distance discriminant analysis model (improved DDA model) of slope hazards caused by blasting vibration in open-pit mining were established by selecting five influencing factors, namely, peak vibration velocity (horizontal and vertical), dominant vibration frequency (horizontal and vertical), and distance from blasting center as discriminant factors. The 24 groups of measured blasting vibration data were collected from 6 field blasting tests, 18 groups of which were used as the learning samples to train the model, and the corresponding discriminant function was established. The misjudgment rate was checked back by the back substitution method, and 6 groups of field data were used as the prediction samples for testing. The results show that the misdiagnosis rate of the DDA model is 5.6%, and the misdiagnosis rate of the improved DDA model weighted by the importance of influencing factors is 0%. The discriminant results of the improved weighted distance discriminant analysis model are completely consistent with the actual results. Because the algorithm is simple, the convergence speed is fast, and the prediction accuracy is high, the improved DDA model provides a new idea for the discrimination of the damage degree of the slope caused by blasting vibration in open-pit mining. © 2021, Science Press. All right reserved.