Fault Feature Extraction And Diagnosis Of Large Data Rotating Machinery Based On Signal Sparse Decomposition

As the core component of the power transmission system,rotating machinery often works under the harsh conditions of high temperature and overload,which is very easy to induce component failure.In order to ensure the healthy operation of the equipment,it is of great significance to carry out condition monitoring and fault diagnosis for rotating machinery.Based on the excellent performance of sparse decomposition in feature extraction,different sparse decomposition algorithms are used to extract the special diagnostic components of bearing faults.It mainly studies the L₀ norm constraint and L₁ norm constraint problems and different solutions for the two types of problems,namely convex optimization algorithm and greedy algorithm,and improves the defects and shortcomings of traditional algorithms.For the greedy algorithm,firstly,the piecewise weak orthogonal matching tracking algorithm is used to extract the characteristic components in the bearing vibration signal.In view of the shortcomings of the algorithm,the S-type function is introduced,and the secondary selection process of the support atom set is added.The S-type quadratic piecewise weak orthogonal matching tracking algorithm is constructed,and successfully applied to the bearing fault diagnosis.The experimental results show that the optimization algorithm has good performance in maintaining feature sparsity,resisting noise interference,reconstruction accuracy and efficiency.For the convex optimization algorithm,the fast iterative threshold shrinkage algorithm is used to extract the features of the original signal.In practical applications,the reconstruction effect is poor.Considering the intra-group/inter-group sparse characteristics of the bearing fault signal,a locally constrained fast iterative threshold shrinkage algorithm is constructed to enhance the solution of the sparse optimization problem and modify the threshold shrinkage function.Applying the optimization algorithm to bearing fault diagnosis can not only restore the fault features realistically,but also maintain the good sparsity of the features,which verifies the robustness of the algorithm in the strong noise background.In view of the shortcomings of the traditional signal sparse decomposition processing method,based on the sparse coding principle,the sparse self-coder and the de-noising self-coder are reasonably stacked,a stacked sparse de-noising self-coder deep learning diagnosis model is proposed,and the eigenmode feature components obtained by the set empirical mode decomposition algorithm are used for the input of the model,and the multi-classification fault diagnosis of bearings is successfully realized.Finally,the proposed in-depth learning diagnosis model is deployed in the cloud of the Internet of Things.Based on the advanced technologies such as big data storage,high-speed data transmission and fine diagnosis analysis brought by the cloud platform technology of the Internet of Things,a precise,real-time and efficient state monitoring and fault diagnosis framework of the cloud platform of the Internet of Things is constructed,and the constant pressure variable pump test bench and motor bearing are taken as the objects,It realizes online monitoring and remote fault diagnosis of rotating machinery.