Fault Diagnosis For Braking System Of Mine Hoist

Braking system is a key unit of a mine hoist, and it plays an important role during both normal and emergent braking processes. As a result, the braking system has a direct effect on the safety of industrial processes. Long time usage and tough work condition can cause inevitable degradation which sometimes will lead to severe accidents.Under the support of the National High-Tech Research and Development Program of China (863Program) under Grant2012AA06A404, the development of the critical equipment for large-sale mine hoist system, the paper firstly summarizes the main faults of the braking system and the current research situation of fault diagnosis for the braking system. Based on the deficiencies of the existing fault diagnosis techniques, our studies are focused on the following tools and methods.First, AMEsim is chosen as the simulation platform, as it can provide many kinds of mature blocks to construct a braking system simulation model. We can invoke different mechanism models and alter the system parameters to obtain a deep insight of the braking system mechanism. A hydraulic braking system experimental platform, which consists of the disc brake and the data acquisition system, is constructed. The experimental platform can help to reform practical work situation. Then the data can be gathered to validate the proposed fault diagnosis schemes.Second, an online monitoring method for spring stiffness based on DBSCAN cluster algorithm is proposed. The braking lining status is classified by clustering the pressure signals with the DBSCAN algorithm. The period during which the braking lining moves inside the clearance is extracted. Then the spring stiffness can be estimated through regression. The monitoring method is put forward based on the braking mechanism and the signals obtained from the simulation platform, then its efficiency is verified by the experimental platform.Third, a scheme is raised to detect the fault that the piston sticks in the cylinder. Based on the different distribution of the friction resistance, the fault signal is differentiated. The fault has an effect on the residual error. Two parameters are advised to reflect the friction resistance. One is the coefficient of the constant term, and the other is energy ratio of the specific frequency band of the residual error. The singularity of the residual error signal is used to detect the jump of the friction resistance.Fourth, an online diagnosis approach for air trapped in oil and leakage is proposed. A feature signal, which can reflect both faults directly, is derived from the mechanism model of the hydraulic cylinder. Feature parameters extracted from the feature signal are related to the magnitude and the category of the faults. An SVM classifier is chosen to isolate the two kinds of faults. At last, a fault diagnosis scheme is proposed, and simulation results show that the scheme fulfills the function of online fault diagnosis effectively.