基于改进工作流的故障预测系统设计

Design of Fault Prediction System Based on Improved Workflow

为降低实验模型构建和参数配置的复杂性,提出一种基于改进工作流的故障预测系统,以可视化的方式设计、配置、执行和部署故障预测实验。该平台不仅能够使专家更容易地使用各种故障预测算法和预测工具,而且能够使不具有算法基础和经验的非专家也参与到实验模型的构建中。通过某表干炉设备的故障预测实验,验证了该系统架构设计的可行性。

A fault prediction system based on improved workflow was proposed to reduce the complexity of model construction and parameter configuration,which was a workflow-based framework to visually design,execute and deploy machine learning experiments.By using this platform,not only experts can use various fault prediction algorithms and tools more easily,but also non-experts without machine learning knowledge and experience can participate in the construction of machine learning experimental models.The feasibility of the system architecture design was verified through the fault prediction experiment of the surface drying furnace equipment.