论文部分内容阅读
为评估冷弯矩形和方形中空截面(RHSSHS)钢梁的有效转动能力,对其进行研究,基于神经网络(NN)和遗传算法(GEP)建立新的数学模型。为对所提出的公式进行扩展,采用了已有研究文献中的大量试验数据。在NN和GEP模型中使用的数据用于涵盖几何和机械性能的8个参数,如截面的宽度、高度和壁厚,内圆角半径,屈服应力,弹性模量与硬化参数之比,初步硬化应变与屈服应变之比以及剪切长度。试验数据验证了所给公式的正确性,将该公式的计算速度和精确性与已有研究给出的半经验公式进行了比较。所给预测模型证明,NN和GEP方法能够用于预测冷弯RHS-SHS钢梁的有效转动能力。
In order to evaluate the effective rotating capacity of steel beams with cold-formed rectangular and square hollow sections (RHSSHS), a new mathematical model was established based on neural network (NN) and genetic algorithm (GEP). In order to expand the proposed formula, a large amount of experimental data in the existing research literature has been used. The data used in the NN and GEP models are used to cover the eight parameters of the geometric and mechanical properties such as cross-section width, height and wall thickness, fillet radius, yield stress, ratio of elastic modulus to hardening parameters, primary hardening Strain to yield strain ratio and shear length. The experimental data verify the correctness of the given formula, and compare the calculation speed and accuracy of the formula with the semi-empirical formula given by previous studies. The predictive models presented demonstrate that the NN and GEP methods can be used to predict the effective rotational ability of cold-formed RHS-SHS steel beams.