Research On Fault Diagnosis And Prediction Methods For Hydraulic Tube Tester

Steel tube is an essential steel product in various fields of the national economy as one kind of fluid transporting pipeline. According to the standard of American Petroleum Institute (API), steel tubes with different types and grades must be able to bear the corresponding pressures. Pressure test has become the necessary part of quality inspection during the production process of steel tube, which plays an important role in the assessment of steel tube. Hydraulic tube tester is a complex system composed of mechanics, hydraulics and electronic control et al., which is especially designed and produced in order to test the pressure resistance ability of steel tubes. Most of hydraulic tube testers are introduced directly from abroad. Due to the constraint of technology secrecy, the mechanism of hydraulic tube tester hasn't been sufficiently comprehended up to now. In the hydraulic system the diagnosis and exclusion of faults bear great technique difficulty, which tends to cost a lot of personnel, money and production time. In addition, as a complex high-press system, the failure of hydraulic tube tester will not only seriously affect the safety of process operation, but also result in huge loss of personnel and wealth. It is important to predict the future fault information via the online estimation of process operation, which will guide the operator in production decision-making and management in time and thus prevent and reduce the loss by system failure. Therefore, it is significant to study the feasible fault diagnosis and prediction methods and fast judge the fault sources, which will improve the performance of machine detection and guarantee the safe reliability of production.Based on the 3# hydraulic tube tester in Baosteel Tube Branch, which is imported from DEMAG Compony, Germany in 1995, this dissertation further studies its work mechanism and thus developed a series of fault diagnosis and prediction methods focusing on the fault characteristics of this kind of hydraulic tube tester, showing important practical values.1. The mathematical models of boost system and balance system are constructed by further analyzing the mechanism of hydraulic system and. simplifying the main hydraulic system of hydraulic tube tester. Consequently, a simulation model is developed for the main hydraulic system of hydraulic tube tester based on the AMEsim software, which realizes the fault outputs and simulation on dynamics of this hydraulic system. 2. Focusing on the common faults of hydraulic tube tester, a method has been developed based on pressure waveform information to discriminate different fault types. Moreover, a 'zooming'knowledge acquisition method is proposed based on the subtractive clustering algorithm in the expert system. The method partitions the boost data of the hydraulic system using the subtractive clustering algorithm and adjusts the parameters in clustering algorithm by communicating with the experts. Therefore, process information can be extracted as much as possible for knowledge acquisition, which realizes the fault types diagnosis in the expert system.3. Based on the diagnosis on fault types, according to the stage characteristics and data-rich property, a stage-based recursive multiway fisher discriminant analysis method is developed for hydraulic component fault diagnosis. For uneven-length problem, a stage-based header-even method is proposed combined with the characteristics of data sampled in hydrostatic testing process, which will improve the diagnosis precision.4. According to the gradual property of hydraulic components faults over batches, a fault prediction method is developed based on multivariate statistical technique and time series analysis. The statistical values of next batch are estimated by building the autoregressive models of T2 and Q statistics, and then compared with the control limits, so that the gradual faults can be predicted.5. A systematic scheme is described for fault diagnosis in hydraulic tube tester, and further one part of functions has been put into online application.An integrated fault diagnosis system of hydraulic tube tester is constructed based on the above mentioned methods. It firstly employs expert knowledge to identify the fault types quickly and correctly and further judges the fault components based on multivariate statistical technique. Then the fault information is predicted based on fault diagnosis results, which can avoid the process failures and improve the production efficiency. The above methods have all been verified in the simulation experiments and some of them have been realized online.