Research For Fault Diagnosis And Simulation Of Turbine Blades

Gas turbine as a representative of the modern power plant is a typical complicatedmechanical system for ships and warships. It is complicated in structure, and its workingcondition is very bad. Above all it is prone to all kinds of mechanical failures. That thecomponent failure of the engine pneumatic, the vibration of the rotating parts,the frayed afterfriction and other kinds of failures is affect seriously the operation of the safety, reliability andefficiency. Therefore, it has the very vital significance that improving and perfecting thecondition monitoring and failures diagnosis technology of the engine can provide a powerfultechnical support. Because the breakdown maintenance manual of traditional paper equipmentexist disadvantages that are carried inconveniently、consulted trivially、storaged difficultly andalternated badly, the appearance of Interactive Electronic Technical Manual (InteractiveElectronic Technical Manual, IETM) overcomes the disadvantages of traditional paperequipment, and becomes the mainstream of the modern equipment support technology inrecent years. Not only can IETM translate paper description, image technology securityinformation into electronic video, audio and image information, but also can locate failuresthrough the interactive way. So the combination of maintenance technology of turbine andIETM has become one of the important topics for the development and application of turbine.This paper describes the IETM system development and key technologies and put theturbine blade fault diagnosis process as an example to make turbine blade Fault Tree, andaccording to the fault tree structure edits fault isolation Data Module. Finally it realized thehuman-computer interactive interface. This system used IE browser as the IETM readersystem to processing various technical information for management and operation. Throughthe human-computer interaction interface maintenance personnel can efficiently and quicklylocate the turbine fault, and make the turbine daily maintenance easier, and the machinepotential hidden trouble can be found and processed urgently. This is to avoid the occurrenceof failure, and can prolong the service life of equipment.This paper researches the turbine blade temperature characteristics and the segmentationmethod of high pressure turbine blade temperature. Simulate the front airway and rear edgeairway fault temperature data with the Kaiser Window function based on analysis of the characteristics of a single blade temperature data. Although the simulated fault data can’tcompletely express the real turbine blade failure data characteristics, it can show a part of thefailure data characteristics, and it has a very good recognition degree compared with thenormal data. It is enough for the simulation data to verify the feasibility of IETM system. Sortthe data firstly before inputting the turbine basic fault information in the IETM system. Wheninput a set of data, the maintenance personnel need judge which is the fault data or which isthe normal data and the fault data match with what kind of fault types also. Classify the inputdata of the blade temperature sample with the LVQ Neural Networks in order to distinguishbetween the good blade and bad blade. Classify the fault type with having been confirmedfault data through the SOM Neural Network.