Design Of A Wireless Multi-sensor Information Intelligent Processing System Oriented To Spinning Cone Penetration

Geotechnical engineering investigation is a key technology link of the civil construction.Its accuracy and efficiency will directly affectt the safety and construction period of the civil construction.At present,the main methods of geotechnical investigation are recovery method and in-situ test method.Among them,the recovery method is easy to cause the stress release or even damage of the sample during the process of drilling,transport:ing and cutting,which results the accuracy of the survey data being reduced and affects the accuracy of the soil mechanics analysis.The cone penetration test of the in-situ test method relatives to the recovery method,it has the advantage of high efficiency,low cost and wide application range.Due to complicated operation of equipment,high failure rate and shallow depth of penetration,the traditional cable-cone penetration test has difficulty in meeting large-scale public facilities and high-rises.The rotary penetration sounding has a deep survey depth and its application prospect is good.But the existing soil mechanics theory research can not give the soil mechanics analysis model applied to the rotary penetration sounding.In tackling these questions,this paper designs the lower computer embedded system which has the functions of multi-sensor acquisition and wireless long-distance transmission.This paper also designs the host computer signal acquisition and processing system.Then it set up the geotechnical classification mapping model of the rotary penetration sounding.(1)The rotary penetration sounding multi-sensor acquisition embedded system design:Used STM32 as the main controller design underground embedded system and ground embedded system.The underground embedded system can measure multi-information such as cone tip resistance,sidewall resistance,underground temperature,humidity,and probe tilt angle.The ground embedded system can measure the penetration depth,ambient temperature and humidity.Based on wireless transmission,The system can achieve the function of real-time interaction,analysis and processing(2)Information processing system design:Used the Lab VIEW virtual instrument design the upper computer information processing system.The system can collect and process the parameter measured by the lower machine.The system can achieve the function of realizing the communication with the multi-sensor acquisition embedded system of the lower computer.And the system can save the test data of the lower computer in a common format for use in subsequent analysis.(3)Based on the optimized support vector machine algorithm,a geotechnical model of spinning and exploratory geotechnical identification is established:The particle swarm optimization support vector machine(PSO-SVM)is used to establish the experimental data mapping model of spinning and penetration test,and the soil mechanics analysis programmer based on artificial intelligence is written.The program implements a mapping from measured cone tip resistance,sidewall resistance,subsurface temperature and humidity,and probe tilt angle to geotechnical categories.(4)The comparison test of 711 sets of data of 5 touch probe holes shows that the classification accuracy of the map model can reach more than 96%,and the validity of the PSO-SVM based spin probe identification model is verified.