An analysis of local stability of soft rock cinder embankment based on the cusp catastrophe theory

In order to make full use of insitu material resources and achieve cost savings, this paper analyzes the engineering characteristics of soft rock waste slag based on screening test, compaction test and CBR(Califonia Bearing Ratio Test)test. Based on subgrade settlement monitoring and by using point mutation, the stability of the roadbed after the foundation is theoretically analyzed, and the roadbed deformation is predicted with the PSO-ELM (Particle Swarm Optimization for Limit Extreme Learning Machine) model to evaluate the development trend of the roadbed stability. The results show that the Chengwu high speed soft rock slag has good gradation, and the CBR value under the optimal water content is in line with the requirements of embankment design, indicating the feasibility of soft rock slag as subgrade filler. The field monitoring results show that the maximum settlement of No. I monitoring section is 71.3 mm, the maximum settlement of No. II monitoring section is 61.4 mm. In the holistic stability analysis of cusp catastrophe theory, the mutation characteristic values of the two monitoring points are all greater than zero and the subgrade is in a stable state. At the same time, in the stage of stability analysis, the sudden change characteristic values of different stages are greater than zero, and with the increasing of the settling time, the sudden change characteristic value decreases, and a way out is obtained. The foundation is also in a stable state in different settling stages. Through the deformation prediction, the two monitoring points are of high prediction accuracy. The applicability of the PSO-ELM model in the prediction of subgrade deformation is verified, and the development trend of the stability of the subgrade tends to be stable, and the applicability of the soft rock residue in the subgrade filling is verified.