基于CSO-SVM的轴承健康状态评估研究

Health Status Identification of Bearing Based on CSO-SVM

摘要:

支持向量机是基于统计学理论的机器学习算法，在解决高维、局部极值和结构选择问题中具有优势，广泛应用于数据发掘中。但是其核宽度和惩罚因子的选择直接关系到支持向量机分类结果。针对上述问题，可采取优化算法对该参数进行优化，达到提高支持向量机的分类精度的目的。鸡群优化算法是近年新提出来的一种全局优化算法，具备结构清晰，全局搜索能力优等优点，在优化问题中得到广泛应用。基于此，提出一种基于鸡群优化的支持向量机模型（CSO-SVM）的健康状态评估方法，并应用在轴承健康状态评估领域中。结果表明，基于CSO-SVM的轴承健康状态评估精度达到97%，明显优于基于传统机器学习模型的健康状态模型的评估精度，具有更好的健康状态识别效果。

Abstract:

Abstract：Abstract: Support vector machine is a machine learning algorithm based on statistical theory. It has advantages in solving high-dimensional, local extremum and structure selection problems. It is widely used in machine learning and data mining. However, the selection of its kernel width and penalty factor is directly related to the classification results of support vector machine. To solve the above problems, using the optimization algorithm to optimize the parameters can improve the classification accuracy of support vector machine to a certain extent. Chicken swarm optimization algorithm is a new global optimization algorithm proposed in recent years. It has clear structure and excellent global search ability. It is widely used in optimization problems. Based on this, a support vector machine model based on chicken swarm optimization (CSO-SVM) is proposed and is used to evaluate the bearing health status. The results show that the accuracy of bearing health state evaluation based on CSO-SVM is 97%, which is much higher than that of the health state model based on traditional machine learning model, and has better health state recognition effect.