基于有限混合分布建模方法的道路轴载模型研究

车辆轴载谱是道路路面设计的关键。对此，提出采用有限混合分布建模方法，精确模拟车辆轴载的分布特性，建立了基础经典分布选择、分布参数校核和基于最小二乘误差优化等方法。并基于四条高速公路的实测WIM数据，校验了该有限混合分布模型，获得车辆单轴和联轴的典型荷载分布模式。结果表明:采用对数正态分布与正态分布线形组合的有限混合分布模型，能很好地拟合车辆轴载的分布特征，R2均在0.96以上，拟合精度高。该混合分布模型具有参数易校验、连续可微等优点，可广泛适用于任意数据类型，其形成的全概率模型也可为道路路面设计提供理论依据。

Study on Axle Load Model of Road Based on Finite Mixed Distribution Modeling Method

Axle load spectrum is of critical importance for road pavement design. In this paper, finite mixed distribution modeling method is proposed to precisely modelling the distribution features of traffic axle loads and the related approaches of foundation distribution selection, parameters calibration and error optimization based on least square method are demonstrated. Several site-specific WIM data is utilized to validate the finite mixed distribution model and typical distribution patterns of single and coupling axle loads are established. The results indicate the finite linear combination of normal and lognormal distributions give a good fit on the WIM data, the calculated R2 are all over 0.96, which show a high fitting precision. The proposed finite mixed distribution models have the advantages of easy parameter-calibration and continuously differentiable, which could be widely applied to any presented traffic axle loads. Moreover, the full probabilistic models from the method can be potentially as the theoretical basis for pavement design or assessment.