基于PSO-SVM的钢筋混凝土梁火灾损伤识别方法

FIRE DAMAGE IDENTIFICATION OF REINFORCED CONCRETE BEAMS BASED ON PSO-SVM ALGORITHM

将粒子群优化(PSO)算法与支持向量机(SVM)相结合,应用于钢筋混凝土梁受火损伤程度识别研究中。输入参数为频率和振型的组合,输出参数为受火时间,构建基于PSO-SVM的损伤预测模型,用简支梁数值模拟验证了该方法的有效性,并与相同条件下构建的SVM损伤预测模型进行对比,结果表明PSO-SVM识别结果准确率更高,适用于实际工程。

The particle swarm optimization(PSO)algorithm and support vector machine(SVM)were used to determine the fire damage degree of reinforced concrete beams.The combined parameters of frequency and vibration mode were taken as input and the corresponding fire time as output,a damage prediction model based on PSO-SVM was established.The effectiveness of this method was verified by numerical simulation of simply supported beam,and compared with the SVM damage prediction model constructed under the same conditions,the results showed that PSO-SVM recognition results were more accurate and suitable for practical engineering.