Research On Fault Diagnosis Of Hydraulic Drive System For New Spindle-type Cotton Picker

With the increase of cotton production scale and the enhancement of mechanized cotton picking,the research and development of new cotton pickers at home and abroad has become a hot trend nowadays.As the core component of cotton picking technology,the research of its driving system is the key technology for cotton picker research and development.Compared with the traditional mechanical drive system,the hydraulic system is widely used in the driving of cotton picking heads and is an irreplaceable part of its power and control because it has the characteristics of large load carrying capacity,small size,light weight,smooth transmission,wide range of realization of stepless speed regulation and fast dynamic response.However,once it fails,it will affect the working efficiency of the cotton picking machine.In this paper,the fault diagnosis of hydraulic drive system for cotton picking head is studied.Based on the horizontal scientific research project "new design,new technology and development of cotton picker ",a new type of ingot picker is developed for the middle and small farmers,which is based on the platform of the hydraulic system test bed of the cotton picker set up by the laboratory.Considering that the driving system has the characteristics of complex nonlinearity and fault diversity,the simulation software,AMESim with low cost and high efficiency,is used to simulate the working condition.After defining the relationship between the model parameters of the hydraulic system and the fault mechanism,the corresponding fault is simulated and the fault samples are obtained through the fault introduction.In order to diagnose and classify the faults of the cotton picking head driving system,considering the fact that the fault data has small sample and nonlinear characteristics,this paper proposes a fault diagnosis method based on the cuckoo algorithm optimization for support vector machine,in which the cuckoo algorithm optimizes the parameter of the support vector machine kernel function.Under the multi-scale training sets and test sets,the diagnosis results of CS-SVM method is compared with the SVM and PSO-SVM methods.The results show that the CS algorithm has a very good optimization effect on the kernel function parameter model of SVM,and the classification accuracy of the fault is also high and the highest correct rate reaches 96.67%.Finally,the preventive measures and methods are put forward for the common fault characteristics of the hydraulic driving system of the cotton picking head,mainly from the oil viscosity control,the size of the gap and the installation of the oil pipeline,and the problem of how to make the system efficient and safe is analyzed in detail.