Dynamic Strain Inversion Of Blade Based On Digital Twin

Such as aero-engine,steam turbine,gas turbine and other turbine machinery widely exist in our industrial production.Most of these equipment are expensive and have many technical difficulties,which will cause great losses in case of failure.Blade is the key component of energy conversion in turbomachinery.Its operation condition is complex.It is easy to vibrate under centrifugal load,aerodynamic load,thermal load and other loads,and then fatigue,crack and other failures occur.Due to the important role of blade in the equipment and its complex characteristics,it is very important for the safe operation of these turbine equipment to obtain the dynamic strain and stress level of blade in real time to evaluate its health status,and finally make early warning and adjustment before failure.On the basis of blade tip timing tip amplitude measurement,combined with the relationship between tip modal displacement and characteristic point modal strain,the blade dynamic strain can be retrieved in real time.Digital twin is the further development and application of digital information technology.It is a technology across many disciplines(including mathematics,physics,computer,machinery,sensors,etc.),which aims to predict and guide the development of the physical world by creating the digital twin of the physical world.In this paper,based on the concept of digital twin,the non-contact measurement of blade dynamic strain is studied.1.Based on the modal analysis theory and the finite element numerical method,the blade mode calculation program(b-solo)is developed to simulate the frequency,modal displacement and modal strain.The calculation results show that the maximum relative errors of frequency rate and modal displacement are 0.06%and 1.17%respectively,So the developed modal calculation program is reliable.Considering the large scale of finite element computing,this paper introduces sparse matrix storage technology into finite element computing,and makes the program more practical by optimizing memory occupation and access.The experiment of stiffness matrix generation shows that the number of sparse storage elements is less than 3 orders of magnitude than the general storage,and the superiority will be more obvious with the increase of scale.So it is necessary to use sparse matrix in the development of finite element program.2.The resonance test of a certain aeroengine blade is carried out.The frequency,displacement,strain and other data of the blade are obtained during resonance,and the strain-displacement coefficient is calculated.The modal displacement and modal strain of blade are calculated by using the developed program b-solo,and the strain-displacement coefficient at the same position as the experiment is calculated.The experimental and simulation results are compared and error analysis is carried out.The results show that the error of each measuring point is within 5%,and the calibration result is of high precision,The feasibility and effectiveness of the dynamic strain inversion based on modal calculation are verified.3.The dynamic strain inversion of blade based on digital twin is studied.Firstly,the application of fundamental mistuning model(FMM)in blade dynamic strain inversion is theoretically deduced,and then the frequency,displacement,strain and other modal characteristics of mistuning bladed disk are accurately predicted by numerical simulation(the maximum error of frequency is 0.15%).At the same time,the influence of detuning characteristics on dynamic strain inversion of blade is discussed.The results show that the variation of strain displacement coefficient of blade is more than 10%,so the accuracy of dynamic strain inversion will be improved by continuously updating the modal characteristics of blade during on-line monitoring,but the sensitivity of tip deformation to axial position is almost unchanged,which does not affect the layout of sensors.