Climate Effection Analysis And Water-damage Disaster Forecast Of Expansive Soils Roadbed Based On Rough Sets Theory

Expansive soil is a high plastic clay with high dispersive grains, and swelling owing to absorbed water, shrinkage due to lost water. The characteristics is extremely harmful on engineering. The expansive soils roadbed is destroyed frequently by rain or dry which is an importart engineering road-block. This dissertation generalized the previous ,achievements, and reviewed:the advanced development of water- destroyed research in expansive soil. Water-destroyed mechanism and predicting calamity of expansive soil soil were studied thoroughly by theory analysis, laboratory tests, large-scale model test, Rough Sets and Neural Net forecast methods, etc.. Some original conclusion were obtained as follows:1. Applying the Rough Sets Theory to classified the expansive soils was presented. Based on the Rough Sets Theory and it's analysis method, the expansive soils classified indexes were arrayed according to their significance and the relating or invalid indexes were eliminated in this paper. There are many indexes and methodes of evaluating the rank of the expansive soils now, but those appraisement result only part depended on some test indexes by calculating rough dependability between index and result. The calculation result indicated the significance of calculating indexes can be supplied to the new test index to optimize expansive soils classification index system and exactly classify the expansive soils.Those data will be .incompatibility and information imperfect due to the expansive soils was discreted by Rough Sets. The Bayes Method was .presented to calculate the conditional probability of the reliability and supportability. And then, those rules should be preserved whose conditional probability is bigger than a given threshold value. The rule of expansive soils classifying generated by logic conjunction and disjunction to the preserved rules. The algorithms optimize the data mining methods which deal with similarity problem.2. The infection of climate to expansive soils roadbed was analyzed. Especially the dry expansive soils has strong expansibility potentially. This is disadvantage to the stabilization of expansive soils roadbed. Via a series of room test and large-scale model test of the medium expansive soils sampled from Ningming segment in Guangxi Nan(ning)-You(yiguan) speedway and weak expansive soils sampled from Cili segment in Hunan Chang(de)-Zhang(jiajie) speedway, the moisture content change rule of the roadbed and the infection of expansive soils during the drying and wetting cycle were deeply discussed under the different climate. The stabilization affected by the rain infiltration and the soil airslaked was analyzed. Because the cranny of expansive soils roadbed, the rain's erosion and overland flow must not be ignored. The design method of the expansive soils roadbed was presented after analysis the structure's mechanical characteristic.3. Probability of the rain and drought which gravely affect the stabilization of expansive soils roadbed were analyzed. Based on one area daily rainfall, the analysis model of Pearson Typeâ…¢and Gumbel distributing were established. The Pearson Typeâ…¢distributing was proved better which can be rather imitated the daily rainfall extremum than Gumbel distributing after goodness of fit. The influence of drought to expansive soils roadbed was also analyzed. The fine-rain translation matrix based on Markov chain was established. And the Feng Lihua's drought model was introduced to evaluate the drought scale. This bring warranty to forecast the water-destroyed of expansive soils roadbed.4. The water-destroyed of expansive soils roadbed was classified by the disaster theory. The water-destroyed of expansive soils roadbed is one of serious geological disaster. The geological disaster character of expansive soils was analyzed according to the attribute character and grade classified of the expansive soils. The analysis indicates the water-destroyed of expansive soils be paroxysmal geological disaster. The risk grade and disaster grade were ascertained. Combining with the routine geological disaster situation evaluating, the character of the expansive soils roadbed evaluation and forecast was analyzed, and the method of evaluation and forecast of the water-destroyed expansive soils roadbed was presented.5. The water-destroyed of the expansive soils roadbed was forecast by Rough-Neural Net. The water-destroyed of expansive soils roadbed forecast is a typical non-linear matter. The improved BP Neural Net was applied. The Rough Rets was applied to analyze training data. The import nodes number was predigested by data predigestion. The nodes number of latent layer was ascertained according to the numbers of rule which confirmed. by Rough Sets data mining.Because there are different outputs for same input in a trained Neural Net by MATLAB. The Monte-Carlo method was presented to simulate the probability of swelling and shrinkage grade, rainfall, drought continuance and roadbed gradient. Those probability were inputted to the trained Neural Net. The statistical result of the output was the water-destroyed probability of the expansive soils roadbed. The calculating result indicates this arithmetic is a patulous arithmetic which can calculate water-destroyed probability of the expansive soils roadbed in any area.