| Integrated Supervision & Control System (ISCS) in metro monitors all devices of the subway and is the substantial guarantee of the reliable operation of the metro, whether it can operate effectively directly relates to the subway's security and stabilization, which determines the high reliability requirements of the ISCS in metro. The reliability of the systemis a problem that calls for immediate solution. However, the higher performance of the ISCS in metro is accompanied by the complexity of structure and function of the system. The complexity of structure includes the extensive use of the hardware/software structure and the redundant structure, and the complex topologies of the backbone network, etc.; while the complexity of function means the system can have many functions. Absolutely, the complexity of structure and function of the ISCS in metro makes the reliability analysis of the system difficult. Therefore, the research of the reliability evaluation method and the analysis of the reliability of ISCS in metro system have significant academic and application value. They can guide the design and maintenance of ISCS in metro, and lay a basis for the further perfection of the system.Based on the analysis of the whole structure of the ISCS in metro, the impact on its subsystem connecting to it, hardware/software structure, redundant structure and the fuzziness of the system, its reliability models were established. And then, the mean value or the stationary reliability indices, the transient reliability indices and the fuzzy reliability indices were obtained by solving the reliability models.Considering the interactions between hardware and software components of workstations, servers and Front End Processor (FEP) for ISCS, different redundancy modes of subsystems devices, and the overall structure of the system, the reliability of the system was analyzed. Meanwhile, with regard to the state space explosion and the readability of Stochastic Reward Net (SRN) for the reliability analysis of complex system, the Hierarchical SRN (HSRN) of the reliability model for the ISCS was established and the equivalence of layering process was proved. This model was analyzed and calculated by Continuous Time Markov Chain (CTMC). The HSRN model can make the mutual function relations clearly and reduce the state space of reliability model.Considering the network topology of communication subsystem for ISCS in metro is double rings, and the increase in the number of nodes and links of the network will result in the exponential explosion in calculation of reliability indices, it goes beyond the capability of the computer to solve the reliability model of network with the normal analytical method. Based on this fact, a hybrid method of reliability analysis for communication subsystem is presented on basis of Dynamic Fault Tree (DFT) and in combination with Monte Carlo method. DFT is used to establish the reliability model of the communication subsystem, and the reliability indices can be obtained by Monte Carlo method based on status arrays to solve the reliability model. This method avoids the state space problems of solving the complex models and provides an effective method for reliability evaluation for the backbone network for ISCS in metro.To study the effect of ISCS in metro on the reliability of subsystems connecting to it, and take the effect of whether the controlled object will continue working on the reliability of controlled object into consideration when the controller unit fails, the concepts of state dependency and functional dependency were introduced. Meanwhile, the modeling process of reliability based on Kronecker algebra and considering the state dependency and functional dependency, and the computational method about reliability indices of repairable system were introduced. Based on the decomposable property of Kronecker-product, the state transition matrix can be divided into two matrices, one is about failure and the other is about maintenance. And then, the state transition matrix about maintenance considering state dependency can be constructed by the method of'stress'matrix, the state transition matrix of system considering functional dependency can be constructed by operational rules. Then, reliability of basic system structures considering failure by state dependency and maintenance by functional dependency was studied, and the reliability model which took the controlled object in the subsystem as the study object was established. This method can construct the state transition matrix, which was used to analyze the reliability of system, avoid the graphic modeling procedure and can be easy to program in computer.The relevant mean value or the stationary reliability indices of the ISCS in metro can be obtained by above methods, however, these methods are somewhat limited in their applications to obtain the reliability measure associated with time. So, the transient reliability indices of system can be obtained by solving these models, which were established by Markov process. The reliability model of hot stand-by redundant system and combined hardware/software system, which are frequently used in ISCS, were established. The reliability model of hot stand-by redundant system considering the system influenced by the effects of failure checking cover ratio and common cause failure was established, and the relations between failure rate, repairing rate, failure checking cover ratio, common cause failure and system state probability, the time for hot stand-by redundant system to get into the stable working state, are discussed. Then, considering characteristic of hardware, software and mutual function relations between them, a reliability model of combined hardware/software system based on Markov process is proposed. This model divides system failures into three categories:hardware failures, software failures and hardware/software interaction failures. Because of the greater number of system state in practical application, the Markov model may cause analysis of system to be very difficult, so the solution based on the recursive network method is presented to solve state transition equation, and thus instantaneous and stable probability of each state are obtained conveniently.About fuzzy reliability of ISCS in metro, considering the system is with complex functions, and which couldn't be described by absolutely normal state or absolutely fault state. In order to define the fair state of the ISCS in metro, the state should be multi state between the absolutely normal state or absolutely fault state. Then, the fuzzy normal state and fuzzy fault state of the ISCS in metro were defined, and the reliability of which was analyzed by the reliability theory about fuzzy state. The reliability indices compared to the conventional indices revealed that the fuzzy reliability could reflect the fault degree of system to the reliability. Sometimes, we can't have the exact values of the reliability parameters but with some uncertainty about these values, then, the fuzzy reliability model was established by affine arithmetic and fuzzy Markov process, and the reliability model was solved by the recursive network method. This method provided a new way for fuzzy reliability of repairable system.From the study above, the dissertation finally forms a reliability analysis frame for the ISCS in metro to obtain the indices including the mean value or stationary reliability indices, the transient reliability indices and the fuzzy reliability indices.This dissertation is supported by National Natural Science Foundation of China: 'Research on the theory and method for evaluating the reliability and safety of integrated supervision & control system in metro'(No.50878188,2009-2011).
|
PDF Full Text Request