Based On FOA-GRNN Long Spiral Bored Cast-in-place Concrete Pile Quality Prediction

Auger Bored Cast-with-pressure Concrete Pile is a new type of pile widely used in the construction of pile foundation in recent years.It is a variable cross-section pile formed by high-pressure pump conveying concrete through the compaction and infiltration of the soil around the pile.How to determine the quality of the pile in the project is the most important problem that engineers need to solve.Firstly,the paper discusses the conventional method and predictive method for the quality inspection of pile foundations,and then studies the bearing mechanism,load transmission and failure modes of Auger Bored Cast-with-pressure Concrete Piles,and also analyzes a variety of factors about effects of the quality of Auger Bored Cast-withtheoretically,including soil layers,piles,construction techniques,time and space effects,etc.In accordance with the results of the analysis,an index system was established to influence the quality factors.Second,because the GRNN has better nonlinear fitting ability and has strong advantages in learning speed and regression ability,it is more suitable for predictive analysis.But the adjustment parameter-smooth factor is difficult to determine.This paper is based on the understanding of the quality of Auger Bored Cast-with-pressure Concrete Piles and GRNN.At the same time,FOA was introduced to optimize the algorithm,so that the optimal parameters were obtained and Quality Prediction Model which involves a total of 9 factors including the construction of the pile body,soil layer,construction,etc.as input layer parameters was established based on FOA-GRNN to predict the quality of Auger Bored Cast-with-pressure Concrete Piles.In the end,the data of Auger Bored Cast-with-pressure Concrete Piles collected in Handan area were applied to the FOA-GRNN model for training.At the same time,predictive analysis was performed and the result compared with BP neural network and unoptimized GRNN model.The results show that although all three prediction models are feasible,the prediction model of Auger Bored Cast-with-pressure Concrete Piles based on FOA-GRNN has more advantages of fewer adjustment parameters,faster convergence,higher precision,and less trapping local minima than the GRNN prediction model,formula method and BP prediction model.It provides a new method for predicting the quality of Auger Bored Cast-with-pressure Concrete Piles.