Yanling Qian

A novel cell mutual detection method for bio-inspired electronic array

Bio-inspired electronic arrays (BIEAs) are circuits that consist of electronic cells. Their self-repairing ability results in high reliability, which brings them many potential demands in aviation field. Recent developments on self-organization and self-repairing methods of BIEA have led to a demand for fault detection as it is the premise of self-repairing. However, the conventional fault detection methods like redundancy and built-in test are not suitable for BIEAs. The hardware consumption of redundancy method is considerably high, and the structure of built-in test is too complex with a large hardware consumption as well. To solve this problem, a fault detection method based on cell mutual detection is proposed in this paper. Referring to the stuck-at-fault errors that usually occur in aviation electronic devices, the method takes the advantage of the features that all the components of the cells are the same. The test vectors are input into all the cells simultaneously through bus and the outputs of each cell are connected to some comparators. The outputs of these comparators are connected to cells as feedback through AND gates that can distinct the faulty cells. In this way, the function of cell mutual detection is fulfilled. A simulation experiment based on FIR filter has been done to validate the feasibility of this method.

Data analysis on incomplete failure of ship electromechanical system based on bayesian method

As a result of the influence of man-made or other factors, the recorded fault data in the process of keeping maintenance of ship electromechanical system are often inaccurate with incomplete record information, low-quality data and small size sample. For the incomplete failure data, this paper will analyze the failure data of the certain type of ship electromechanical system thoroughly and comprehensively through Bayesian method. Sorting the data firstly and then analyzing graphical statutory representation to determine the failure trend, I will adopt the homogeneous Poisson process and the non-homogeneous Poisson process (power law process (PLP) and Duane) on quantitative analysis to establish failure data analysis model with Bayesian method and obtain the analysis result through PyMC3 probability programming. Finally, the Bayesian estimation of the average failure frequency of ship electromechanical system will be concluded in this paper.

Information extraction method and its application in Chinese equipment technical manual based on rule-matching

Technical manual is vital source of fault diagnosis and maintenance knowledge. To deal with the mass of manual text, which influence the speed of knowledge acquisition, and then effect the diagnostic efficiency, this paper proposes a method for extracting information of Chinese technical manual text based on rule-matching. With the help of domain ontology and syntactic analysis, extracting rules could be established to extract manual text information, and reasoning algorithm based on hierarchical directed graph is built to assist the fault diagnosis. The information extraction method presented in this paper can supply knowledge to maintenance staff and help to diagnose the fault automatically.

A method of failure analysis for failure modes and effects analysis based on stress environmental model

Computer aided analysis for Failure Modes and Effects Analysis (FMEA) is an important development direction. Failure analysis (failure mode and failure mechanism analysis) is the most critical and the most difficult step in implementation of FMEA. The procedure of failure analysis is lack of formal and standardized method, which hindering the computer integrated process. In face of complex products, the cost of time and manpower is very large and the analysis result is usually inconsistent, redundant and incomplete. Therefore, the accuracy of result cannot be guaranteed. To solve the above problems, this paper proposes a method of failure analysis for FMEA based on stress environmental model. The engine cylinder is used as an example to explain the method. Firstly, the source and classification of stresses are systematically studied and the corresponding stress identification rules is put forward according to operating conditions of product; secondly, stress environmental model of part or component is established based on the relations and constraints among stresses; finally, an analysis method with formal structure is established through analyzing the characteristics of failure mechanism and failure mode, and the results were summarized based on the analysis of engine cylinder. The model based analysis method presented in this paper can effectively ensure the accuracy of analysis result, and the analysis procedure with formal and standardized properties is also convenient for computer integration at the same time.

Research on CBM information system architecture based on multi-dimensional operation and maintenance data

Condition Based Maintenance(CBM) is a preventive maintenance strategy based on the actual running state of equipment. There are some deficiencies in the implementation of traditional CBM method. Firstly, how to acquire data used by CBM reasonably and improve the data coverage of equipment information is a problem to be solved. Secondly, the CBM method is still a maintenance strategy, when facing the different types of data, how to manage these data with characteristics of multi-sources, heterogeneous and different time scales for convenient searching and data processing also need to be studied. At the same time, how to choose a reasonable model to deal with this data, and then quickly give the maintenance strategy to the staff is another problem. Above all, this paper presents an CBM information system architecture based on multi-dimensional operation and maintenance data, including extensible data acquisition method; data management method based on Ontology Modeling and data processing method based on model selection. The architecture proposed sets data acquisition, data management and data analysis in one system, effectively integrating available information and making the whole analysis process becomes simple, and shortening the analysis process and improving the efficiency of data analysis through computer technology and information technology.

Development of configurations for lookup table-based Embryonics using graphic mapping: A case study

This article presents the development of configurations for bio-inspired self-healing cellular arrays known as Embryonics (embryonic electronics). In the Embryonics design, the configurations are employed to define the functionality and connections of each cell. However, developing configurations of the Embryonics is a time-consuming and challenging work due to lack of effective tools. In this article, an approach is proposed to develop configurations using graphic mapping, which also optimizes the length of configurations for the Embryonics. Using metric embedding, the problem of configurations is exactly formulated to binary quadratic assignment problem and routing problem with constraint of the Embryonics architecture. Since binary quadratic assignment problem is nondeterministic polynomial-time hard, a genetic algorithm is used to tackle this problem for achieving high-quality placement. Due to the limitation of communication bandwidth, how to resolve congestion is also an important issue. An improved ant colony algorithm is presented to realize routing of the Embryonics based on the result of placement. Configurations of the Embryonics are formed according to the result of placement and routing. Experimental result on a 4 × 4 multiplier demonstrates that developing configurations for lookup table–based Embryonics using graphic mapping can lower the difficulty of the Embryonics design and optimize placement and routing of Embryonics.

Reliability analysis of ship equipment based on maintenance data

Field maintenance data represents the performance of product better under actual conditions of use and maintenance, which has important significance for the research of spare parts for ships. Currently, failure data are obtained by the way of routine inspection to research the life of spare parts. Because those inspections are not constantly monitoring but timing detection, the failure times cannot be gathered. To solve the difficulty above, a special method for data recording and interval-censored data obtaining is proposed. Then, the EM algorithm is used to estimate the two-parameter Weibull distribution.