According to soft soil foundation physical and mechanical properties, generally considered that its settlement process is ap⁃proximate to an inverse“S”shaped curve. To study soft soil foundation settlement process and predict ultimate settlement, based on above mentioned cognition, comprehensively considered the mathematical nature of soft soil foundation settlement staged development and biological growth model, then selected high adaptable Weibull growth curve model. Using uniqueness of genetic algorithm on geo⁃technical multiparameter processing and nonlinear problems, carried out fitting for 3 different areas representative soft soil foundation project cases settlement observed data. The result has shown that soft soil foundation after loading, its settlement development will be generally experienced occurring, developing and gradually stabilizing 3 stages similar to biological growth;and the inverse“S”shaped growth model can reflect its staged settlement. Using Weibull model can judge corresponding time settlement stage based on bending point position, that helpful to control operation and loading process. Using genetic algorithm can solve nonlinear geotechnical engineer⁃ing back analysis problems better. Taking the residual sum of squares as objective function, according to residuals analysis can know that the Weibull soft soil foundation settlement model from genetic algorithm has higher accuracy.%根据软土地基的物理力学性质，普遍认为其沉降过程近似为反“S”形曲线。为了研究软土地基沉降过程以及预测最终的沉降量，本文运用这一结论，综合考虑了软土地基沉降的阶段性发展与生物成长模型的数学性质，选用了适应性较高的Weibull成长曲线模型，利用遗传算法在处理岩土类多参数以及非线性问题上的独特优势，通过对3个不同地区具有代表性的软土地基所选工程实例的沉降观测数据进行拟合。结果表明：软土地基经过加载后其沉降发展一般会经历一个类似于生物成长规律的发生、发展、逐步稳定的三个阶段，且反“S”形的成长模型能够反映其沉降的阶段性；采用Weibull模型能够根据反弯点的位置来判断对应时刻所处的沉降阶段，有利于控制施工以及加载过程；运用遗传算法能够很好地解决非线性岩土工程反分析问题，以残差平方和作为目标函数，根据残差值分析可知，用遗传算法得到的Weibull软土地基沉降模型具有较高的精度。
展开▼
