针对铁路沉降中的新问题,采用实测沉降数据,运用统计学原理对三点泊松模型、指数模型、对数模型和双曲线模型对沉降变形的拟合效果进行了研究,并推荐了科学的预测评估方法。结果表明,各个模型均可以满足使用要求,且指数模型的拟合效果最佳,双曲线模型次之,对数模型最差。 Aiming at the new problems in railway subsidence, the measured subsidence data are used to study the fitting effect of three-point Poisson model, exponential model, logarithmic model and hyperbolic model on settlement deformation by using statistical principles and to recommend a scientific Prediction method. The results show that each model can meet the requirements, and the exponential model fits the best, followed by the hyperbolic model and the logarithm model.