Research On Health Detection Technology Of The Cantilever Beam Structure Based On PVDF Sensors Array

As a very common structure member,the cantilever beam is widely used in mechanical,aerospace,energy,civil engineering and other fields.In the long term service,the cantilever beam is usually affected by complex load,and it is easy to produce cracks on the interior or surface,and even lead to the occurrence of the major accidents.Therefor,the research on health detection of the cantilever beam is the focus and hot spot in the field of engineering application.In many structure health monitoring methods,damage identification methods by using the natural frequency and other vibration characteristics have been widely used,which have the advantages of fast,efficient and non destructive.However,it is limited only using one of the above vibration characteristic information,whether the real time measurement,the signal feature extraction or the damage identification,which can not meet the actual needs.So,a new two-step method of the structure health monitoring is proposed based on PVDF piezoelectric sensors array combined the vibration characteristic of the cantilever beam.It is divided into two steps: in the first step,the damage number and the damage positions of the cantilever beam are identified using the difference of the normalized strain modal before and after damage;the second step,the damage degree of the cantilever beam is identified using BP neural network model.Firstly,for the cantilever beam structure,the surface crack of the cantilever beam is equivalent to a rotating spring,and the mathematical model of crack cantilever beam is established,the relationship between the difference of the normalized strain modal before and after damage and crack parameters is derived,the number of the crack and the crack position are identified using the relationship;on this basis,the natural frequency changes of the cantilever beam before and after damage are taken as input parameter of the BP neuralnetwork,the damage degree is taken as output parameter,so as to identify damage degree of the cantilever beam.Secondly,we consider the aluminum alloy cantilever beam structure with equal section as the research object,the damage identification of the single and double cracks is simulated,the relationship between the difference of the normalized strain modal before and after damage and crack parameters is obtained,the BP neural network model of the cantilever beam with equal cross section is established at the same time.In order to further verify the effectiveness in practical engineering application of the proposed two-step method in this paper,the damage identification experiment of the single crack and double crack of aluminum alloy cantilever beam is studied by using PVDF piezoelectric sensors array.The identification results using the two-step method in this paper are compared with the identification results using the transfer matrix method.The results show that the proposed two-step method in this paper has the following advantages.Firstly,whether the damage degree identification of the single crack or the damage degree identification of the double crack,the accuracy of the proposed two-step method in this paper is higher than the transfer matrix equation method;Secondly,the principle of the transfer matrix equation is based on mathematical model,so it is only suitable for these structures such as the simple rules and the mathematical equations of the structure can be built,the complex structure are often difficult to establish mathematical models,so the transfer matrix equation method can not identify the complex structure.The proposed two-step method in this paper is separated from the mathematical model,which is applicable both rule structure and complex structure,it has better prospects for engineering application;Thirdly,computation cost of the proposed two-step method in this paper is smaller and time consuming is less.Finally,we take the small wind turbine blade as the research object in order to further study the effectiveness of the proposed two-step method in this paper for the complex structure health monitoring.A small wind turbine blade is designed based on Wilson model,the chord length and twist angle of each blade of small wind turbine blades are corrected using MTALAB optimization function.Simulation modal analysis of blades under differentdamage states are carried out.The corresponding BP neural network model is established by using the extract the modal frequency changes before and after blade damage through modal analysis.Then,experimental modal analysis of small wind turbine blades are conducted using PVDF sensors array.The experiment results showed that the proposed two-step method was also effective for damage identification of the small wind turbine blade.