关键词: 人工神经网络, B-P网络, 沉积物识别, 粒度参数

Abstract: The so-called "old red sands" are semi-cemented fine and medium sands with colors in red or brown red, distributing along the coasts in the southern Fujian and eastern Guangdong provinces, China. In recent ten years, different opinions raised about the origin of the "old red sands". Some people considered that the "old red sands" were coastal deposits with their sources from nearby places, while others believed that they were products deriving from aeolian sands.Artificial neural networks (ANN) are a recently developed information processing technique, which is most likely to be superior to other methods in processing data of nonlinearity and ill-definition and corrupted by significant noise. Several B-P neural networks with three, four, five, six and seven inputs were constructed for identification of the origin of the "old red sand" and each network has one intermediary layer and one output. The inputs of the networks were the grain size parameters of the sediments, such as Mean, Sorting, Skewness and Kurtosis. The grain contents of different sizes from gavel to silt/clay are also used as the inputs for the networks. Modern coastal dune sand and beach sand collected from South China coasts were used to train the B-P networks. The fully trained B-P networks couled recognized most of the "old red sand" as aeolian sand, indicating a close relationship between the "old red sand" and aeolian sand. The result of the identification by the networks with gavel and silt/clay content as the inputs of the networks indicates that it is an effective parameter for identifying aeolian deposits. However, the network only with the gain contents of all sizes from gavel to silt/clay as the inputs fails in identifying of the "old red sand". The result also shows that the "old red sand" has poorer sorting and finer particle size than those of the modern coastal dunes.

Key words: artificial neural networks, back propagation network, sediment discrimination, grain size parameters