Study On Fault Diagnosis Of Nonlinear Systems Based On Permutation Entropy Theory

With the gradual advancement of the industrial age,the connection among the structure,system,and control of mechanical equipment has become increasingly close and complex.No matter which part has a fault or has failed,it will get the whole body and the mechanical equipment into trouble.Performance can't achieve the expected function,the control system fails,and the catastrophic consequences.In addition to causing huge economic and property losses,mechanical equipment failures can cause serious casualties,environmental pollution and adverse social impacts.Therefore,the study on the fault characteristics of mechanical equipment is an important issue to ensure the smooth progress of production and life.The problem diagnosis and research about fault diagnosis are often based on practical problems,and have a strong engineering application background.The actual engineering environment is mostly a complex nonlinear system,which contains almost all kinds of nonlinear factors.Therefore,it is necessary to simulate the actual working environment of the mechanical equipment with a nonlinear system and identify and extract the fault signal characteristics.In this thesis,based on permutation entropy theory,the pseudo-phase plane method and the fractional-order permutation entropy spectrum are used to analyze and study the typical faults of rolling bearings.The main research contents are:(1)A new pseudo-phase plane method based on permutation entropy and its related theory to identify the characteristics of rolling bearing fault signals is studied.Through the traditional permutation entropy theory,the displacement excitation,the primary permutation entropy and the second permutation entropy in the improved method are respectively used as state variables,which constitute the pseudo phase plane forms of displacement excitation-primary entropy and displacement excitation-second permutation entropy.The pseudo phase plane analysis of the simulated signal and the measured fault signal from rolling bearing is carried out.The results show that the pseudo phase plane method can identify the doubling periodic characteristics and other signal features in the vibration signal.(2)Based on the study on the permutation entropy theory of integer-order permutation,the effectiveness of the fractional-order permutation entropy is verified by the simulation signal,and it is verified that the fractional-order permutation entropy has certain sensitivity to the abrupt in the rolling bearing fault signal.Then,based on the fractional-order permutation entropy theory,a new method of fractional-order permutation entropy spectrum is proposed to extract the weak abrupt signal,and the effectiveness of the method is verified by numerical simulation.The results show that compared with the integer-order permutation entropy and the integer-order permutation entropy spectrum,the fractional permutation entropy spectrum is more effective for feature recognition of weak fault signals,and the fault frequency component is more obvious,which provides a feature for fault diagnosis of weak abrupt signals in the future.Finally,the work of this thesis is summarized,and the future research work is prospected.