论文部分内容阅读
我们从1984年起开始采用模式识别方法研究长输管道的土壤腐蚀问题,如:沿线土壤腐蚀性分级和涂层在土壤中的老化速度等,均取得很好效果,现举例说明: 1.长输管道沿线土壤腐蚀性等级的分类 1980-1983年我局对两条长输管线(总长400km)进行土壤腐蚀调查,测试了沿线62个典型土壤点的理化性质,其中还设立了11个埋片失重试验点(16Mn钢,埋设1年半)。采用常规的多元回归等方法处理所得到的数据组,效果不理想;而光凭少量埋片失重数据难以了解沿线土壤腐蚀性全貌。后来,采用主成份分析(PCA)处理,获得一系列有意义的结果:
Since 1984, we have used pattern recognition methods to study the soil erosion of long distance pipelines, such as the soil corrosiveness classification along the soil and the aging rate of the coating in the soil, etc. The results are as follows: 1. Long Classification of Soil Corrosiveness Levels along Pipelines During 1980-1983, we investigated the soil erosion of two long-distance pipelines (total length 400km), tested the physical and chemical properties of 62 typical soil sites along the route, and set up 11 buried films Weight loss test point (16Mn steel, buried one and a half years). Using conventional multiple regression methods to deal with the data set obtained, the effect is not ideal; and light with a small amount of buried weight loss data difficult to understand the corrosive landscape along the soil. Later, using principal component analysis (PCA) processing, a series of meaningful results were obtained: