摘要
针对风电机组运行工况复杂多变,机组状态监测数据量大、多源、复杂、增长迅速等特点,现有的异常预测方法在面对大数据时难以既保证预测精度又进行快速处理,故提出了结合Hadoop批处理技术和BP神经网络的风电机组在线异常预测模型,对设备状态信息进行异常预测。实验结果显示,该异常预测方法在保证精度的前提下具有较好的加速效果,可以为风电场维护人员提供重要的参考信息。
According to the working conditions of wind turbine generator monitoring complex, large amount of data, multisource, complex, the characteristics of rapid growth, the abnormal current prediction methods in the face of big data to ensure accuracy and rapid processing, the proposed combination of Hadoop batch processing technology and BP neural network of wind turbine online anomaly prediction model, abnormal prediction of equipment state information. The experimental results show that the method has good acceleration effect under the premise of ensuring the accuracy, which can provide important reference information for the wind farm maintenance staff.
出处
《电脑知识与技术》
2017年第1X期245-247,共3页
Computer Knowledge and Technology
基金
国家科技支撑计划课题(2015BAF22B00):面向石化冶金行业流程生产过程的工艺软件与知识库研发
上海市信息化发展专项(201403028):面向钢铁工业节能减排的加热炉大数据分析与优化平台
上海市信息化发展专项(201501046):EMC项目远程智能监控中心关键技术研究与建设