您的位置：首页 > 论文页面
基于LS-SVM的油气管线点腐蚀深度扩展行为预测
发表时间：2013-09-30 浏览量：1429 下载量：378
| 全部作者： | 程光旭，李珺，范志超 |
| 作者单位： | 西安交通大学化学工程与技术学院；哈尔滨电气股份有限公司环保事业部；合肥通用机械研究院科研管理部 |
| 摘 要： | 基于最小二乘支持向量机（least squares support vector machine，LS-SVM）建立了油气管线点腐蚀深度时间序列预测模型，即LS-SVM模型。针对LS-SVM模型参数优化过程十分复杂的情况，在标准粒子群优化（particle swarm optimization，PSO）算法的基础上，引入改进粒子的位置初始化、自适应惯性权重和改进速度更新公式这3种新的进化策略，提出一种混合粒子群优化（hybrid particle swarm optimization，HPSO）算法改善PSO算法的参数寻优性能。仿真试验显示：HPSO算法比标准PSO算法具有更强的搜索能力和更高的收敛精度。工程实例计算结果表明：与GM(1，1)模型、自回归移动平均（auto-regressive moving average，ARMA）模型和BP神经网络（BP neural network，BP-NN）模型相比，LS-SVM模型的预测性能更好。 |
| 关 键 词： | 石油、天然气能；油气管线；点腐蚀深度；最小二乘支持向量机；参数优化 |
| Title： | Growth behavior prediction of pitting corrosion depth for oil and gas pipelines based on LS-SVM |
| Author： | CHENG Guangxu, LI Jun, FAN Zhichao |
| Organization： | School of Chemical Engineering and Technology, Xi’an Jiaotong University; Department of Environmental Protection, Harbin Electric Company Limited; Department of Research Management, Hefei General Machinery Research Institute |
| Abstract： | In this paper, a pitting corrosion depth time series prediction model for oil and gas pipelines, namely least squares support vector machine (LS-SVM) model, is established based on LS-SVM. As for the complex situation of the parameter optimization process in LS-SVM model, a hybrid particle swarm optimization (HPSO) algorithm is presented by the introduction of the modified initialization of particle positions, self-adaptive inertia weight and modified velocity updating equation to improve the parameter optimization ability of particle swarm optimization (PSO). The simulation experiment indicates that HPSO algorithm has stronger search ability and higher convergence precision than standard PSO algorithm. The calculated results of an engineering case show that LS-SVM model performs a better prediction performance when comparing with GM(1, 1) model, auto-regression moving average (ARMA) model and BP neural network (BP-NN) model. |
| Key words： | oil and gas energy; oil and gas pipelines; pitting corrosion depth; least squares support vector machine; parameter optimization |
| 发表期数： | 2013年9月第18期 |
| 引用格式： | 程光旭，李珺，范志超. 基于LS-SVM的油气管线点腐蚀深度扩展行为预测[J]. 中国科技论文在线精品论文，2013，6（18）：1699-1708. |
0
评论数 0请您登录
暂无评论
下载全文
多维论文