Methods Of Detection On Moisture Of Railway Subgrage By Using The Ground Penetrating Radar

With the high speed of the train and axle weight increasing, subgrade condition becomes worse. This restricts the development of the railway. We must establish a kind of rapid and efficient evaluation method to detect subgrade condition. Ground penetrating radar technique has advantages of efficient, rapid, high progress and lossless, which in line with the requirements of the railway operation development.This paper uses the railway subgrade detection and evaluation as a starting point. We use the filling of the dielectric properties, radar wave propagation and radar wave attributes as the research object. By means of laboratory test, numerical simulation, physical model test and theoretical analysis, electro magnetism, seismic wave theory, digital signal processing, finite difference time domain (FDTD), statistics and pattern recognition, we make the research of radar wave in loss layered media propagation rules, layered composite material of the dielectric constant of the mathematical model, the radar wave in medium dielectric properties of parameter change response, radar wave on subgrade moisture content changes in response, radar wave attribute extraction and attribute classification recognition problem. Finding the law of the dielectric constant and moisture content of composite materials, we establish the mathematical model of dielectric constant of different materials. We reveal the relationship of the radar wave direct coupled wave, reflection wave, reflected wave attenuation and the waveform characteristics and properties of medium, and the moisture content of the relationship between radar wave energy,100m bandwidth percentage of energy and medium. We propose extraction method proposed attribute parameters of time domain and frequency domain related, and using artificial neural network for railway subgrade moisture containing radar attribute classification recognition and on-site verification. Finding the law of the moisture content of composite materials, establishing the mathematical model of dielectric constant of different materials; revealing the radar wave direct coupled wave, reflection wave, reflected wave attenuation and the waveform characteristics and properties of medium, we propose the relationship of the moisture content between radar wave energy and the energy of 100m bandwidth percentage.We propose the method of time and frequency domain, put forward the attribute parameters related, and using artificial neural network on water properties classification and field validation. To provide theoretical foundation and scientific basis for the detection and evaluation of the status of railway subgrade, it has certain guiding significance to ensure train operation safety. In this paper, the main research results are as follows:1. Through the test chamber on the dielectric property of the commonly used several railway subgrade filler, it revealed the relationship between water content of roadbed filler permittivity, conductivity and subgrade filling rate, degree of compaction. The dielectric constant relationship model of the roadbed filler are as following:(1) The permittivity and conductivity increases with the volumetric water content, and the relationship between the volumetric water content, the dielectric constant, and the conductivity are cubic function(2) The dielectric constant of the subgrade is sensitive with the volume of the moisture instead of the compaction factor(3) According to the principle of minimum variance and multiple linear regression equation, several optimum permittivity models of the railway roadbed filler are established:Improved soil and loess dielectric model:εr2/3= θmεm2/3 +θwεw2/3 +θaεa2/3Silty dielectric constant model:εr1/1.3=θmεm1/1.3+θwεw1/1.3+θaεa1/1.32. The railway embankment different dielectric properties of numerical analysis is performed by using the finite difference time domain technique (FDTD) model, revealing the radar response characteristics of the dielectric properties of the dielectric properties of the radar and deducing mutual relationship between the temporal characteristics and the subgrade phases; a function model of GPR reflection amplitude and subgrade dielectric properties and propagation distance is established.(1) Permittivity response of the radar signal1) The radar waves propagating is directly proportional to the square root of the dielectric constant of the medium;2) The amplitude of the reflected wave is cubic function of the dielectric constan of the filler, the relationship is as following: y1= 0.0154 εr3 -1.1645 εr2 +30.2630 εr +140.02363) The direct coupling wave amplitude is negatively correlated to and the logarithmic of the dielectric constant, and the relation is: A=-595.4lnεr+2715.34) With the concept of the amplitude attenuation, which is 20 times the logarithm of the amplitude of the reflected wave amplitude ratio coupled directly; amplitude attenuation is a logarithmic function of the dielectric constant, the relationship is: y=-6.6755lnεr +27.2685(2) Electrical conductivity only have an effect on reflected wave of radar but have no effect on the propagation time and the directly coupled wave of ground penetrating radar (GPR), indicating that only works on radar attenuation. It shows approximate negative exponential relationship between reflected wave amplitude and electrical conductivity, while attenuation of reflected wave amplitude shows a linear relation with electrical conductivity.(3)Reflected wave amplitude is inversely to the square root of radar wave propagation distance; the mathematical model of reflected wave amplitude and dielectric constant, electrical conductivity and propagation distance was derived:3. The railway embankment water model test reveals the relationship between the state of the direct coupling wave, reflected wave radar energy, energy percentage of megabytes of bandwidth and other parameters and subgrade water status(1) The direct coupling wave amplitude decreases with increasing volumetric water content, both negative exponential correlation; reflection amplitude decreases with increasing volumetric water content, both non-linear relationship; directly coupled wave and reflection wave amplitude ratio is a function of the volume of moisture(2) The total energy of the radar decreases with the volumetric water content, both negative exponential correlation, and the expression is: E = 11482ω-1.302。(3) With the increasing volume of water content, the peak frequency radar has a tendency to move to the low side.(4) According to extracting energy percentage 100MHz bandwidth, the relationship between the 100MHz bandwidth energy and volume percentage of moisture is analyzed, and the results show that the volume of water content and energy part of megabytes of bandwidth have a higher percentage of dependency:1) For single model, the percentage bands energy range, including 200-300MHz, 300-400MHz and 400-500MHz, change significantly with the volumetric water content. and it is linearly related the volumetric water content2) For multi-tier model, the percentage bands energy range, including 0-100MHz and 300-400MHz, change significantly with the volumetric water content, and it is linearly related the volumetric water content3) Taking the 300 MHz as the cutoff point, the percentage of energy mono layers range of 0-300MHz frequency decrease with the increasing volumetric water content, the percentage is greater than the energy 300MHz range increases with the volume of water rate increases; for multi-tier model,0-300MHz band energy range with the percentage of the volume of the water content increases, the percentage of energy greater than 00MHz range with the volumetric water content increases. Therefore, the frequency of 300MHz can be used as roadbed advantage aqueous state(5) The proposed parameters radar, including energy directly coupled wave amplitude, the amplitude of the reflected wave, the reflected wave and the direct coupling wave amplitude ratio (attenuation) percentage band energy and peak frequency, can be used to evaluate the ground penetrating radar property subgrade moisture conditions.4. According to the theory of seismic attributes, this paper clarify the nature of GPR attributes, functions, propose classification and extraction of ground penetrating radar properties. In the time domain analysis method to extract the maximum peak amplitude, the maximum amplitude of the valley, the average peak amplitude, average energy, direct coupling wave amplitude of the reflected wave amplitude, attenuation, mean square amplitude and the amplitude kurtosis 19 kinds of radar attributes in frequency Domain analysis methods, focusing on the power spectrum estimation method to extract the peak frequency, wideband total energy, frequency band energy, energy percentage frequency band specified frequency energy, low-band frequency energy and frequency domain series 13 radar attributes used to evaluate the status of the railway embankment water.5. For the railway embankment water status classification method, the BP neural network classification methods based on the ground-penetrating radar properties is proposed. Model tests were selected from four types of data aqueous state, each 100 group,80 groups of neural networks,20 groups of test, the test results are correct rate of over 90%, to meet the railway embankment water status evaluation requirements. And apply the training stable network of 200 sets of data in central and southern Shanxi channel overload test section to test and verify the test results and the field situation basically. It is feasible to test and evaluate the railway embankment aqueous state by using the Ground Penetrating Radar attributes and BP neural network.