Failure prognostic system of equipment has great significance to realizeequipment early failure forecast, guarantee the equipment safe long-term operation,reduce the whole life cycle cost of equipment, improve equipment utilization, andimprove equipment management level and the implementation of scientificmaintenance.The basic principles of similarity-based modeling are introduced, a dynamicsuper ball model are established by filtering out the normal historical operating dataof a300MW unit primary fan for boiler. The predicted value is calculated. Similaritycurve of Primary Fan for boiler and similarity curve of all the points of primary fanfor boiler are got by using Euclidean Distance measure method and similaritycalculation formula of measurement of points, And the sensitivity of each measuringpoint was calculated, and showed the device’s common troubleshooting table.Through similarity coefficient measure method, each point similarity value within thenormal running time is calculated, and made the minimum similarity value as similarbaseline. Finally, development visual interface of the boiler primary fan failureprognostic system based on similar modeling, which based on the Windows2003Server operating system, using ASP.NET realized connection between Web serverand database, using SQL Server2005system as back-end database, using VisualStudio2010. Net system as development platform and the system are testing.Studies show that the failure prognostic system can make the right warning toprimary fan running state, equipment similarity curves calculated by using EuclideanDistance measure method are reliable and reasonable, primary fan for boiler warningWeb client can make timely and accurate warning. |