Development of ANFIS frameworks to assess the stiffness modulus of soft ground soil in cold area
摘要
Because of soft ground soils unique properties—Like a low permeability, high water content, poor bearing capacity and a high void ratio—soft soils pose serious obstacles to the development of infrastructure and the environment. One of the most essential factors when constructing geo-structures is the stiffness modulus (Gs) of soft ground soils. The present tendencies in sustainable development, on the other hand, advocate for the employment of less costly technologies. The calculation was performed using the ANFIS (Adaptive Neuro-Fuzzy Inference System). Its reliability is greatly impacted by the ANFIS hyperparameters, which must be chosen utilizing metaheuristic optimization techniques. This is accomplished via the employment of the Pufferfish Optimization (PO) and the Walrus Optimization (WO). Six distinct input factors have an impact on the stiffness modulus of soft ground soils (Gs), according to a computer database that includes 197 test results from previous studies. During the framework’s creation and evaluation, 75% of the data was utilized from literature as a learning set as well as 25% as an evaluation set. The evidence reveals that both ANFIS-Wa and ANFIS-Pu have a significant likelihood of accurately determining the value of Gs. Every model has been determined to be dependable and reliable based on the assessment criteria, logical reasoning, and score rating. The ANFIS-Wa model performs somewhat better than the other model regarding the goal.