Research On Fault Diagnosis Of Hydraulic AGC System Based On Information Fusion

Hydraulic AGC system has many advantages,such as fast response,high rolling accuracy,variable mill equivalent stiffness,small volume,light weight,high power and so on,so it is widely used in the field of cold rolling.But because it is a kind of mechatronics and hydraulic integration equipment,and the working environment of steel plant is relatively bad,so the hydraulic AGC system will inevitably have some faults.This paper takes the hydraulic ACG system of a cold rolling production line of a company as the research object,mainly studies the fault diagnosis and rapid positioning of common faults in the hydraulic AGC system.Aiming at the problem that hydraulic AGC lacks fault samples and is difficult to carry out field experiments,we will obtain fault data through accurate modeling and fault simulation,and then use the information fusion fault diagnosis method based on wavelet transform and support vector machine to diagnose and classify the common faults of hydraulic AGC system.The specific work contents are as follows:(1)Firstly,the control principle of the hydraulic AGC system is analyzed,and the mathematical model of the system is established according to the control principle.Combined with the hydraulic circuit principle of the AGC system,the mathematical model of each important component in the AGC system transfer function is replaced by the physical simulation model,and the complete physical simulation model of the hydraulic AGC system is established.The accuracy of the model is verified and analyzed.(2)Through literature review,the fault types and fault mechanism of hydraulic AGC system are summarized and analyzed.Taking the frequent faults in Baosteel cold rolling plant as an example,the fault simulation of hydraulic AGC system based on simulation model is carried out,and the pressure,displacement and flow signals obtained from fault simulation are analyzed.It is found that the fault characteristics obtained from fault simulation match the actual fault,The feasibility of fault simulation is proved.(3)The wavelet transform is used to decompose the pressure and flow signals,and the high and low frequency pressure coefficients obtained from the signal decomposition are processed numerically.Several fault characteristic parameters with the most obvious change trend are found when the fault occurs.Through further research,the diagnosis effect of fault characteristic parameters for different degrees of single fault and different faults is analyzed.(4)The information fusion fault diagnosis method based on wavelet transform and support vector machine is used to classify the faults.The OVO SVMs model is trained with the fault feature parameters extracted by wavelet transform as input and fault type as output.The parameters of CV-SVM are optimized by grid search and PSO algorithm,and the optimization results are compared.Using the constructed sample set to train and test the classification model,and the diagnosis effect of the model is analyzed.