Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-09-28
2001-05-01
Wright, Norman M. (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S038110, C714S048000, C703S014000, C703S017000, C703S022000
Reexamination Certificate
active
06226760
ABSTRACT:
BACKGROUND AND SUMMARY OF THE INVENTION
This application claims the priority of 197 42 466.5, filed Sep. 26, 1997, the disclosure of which is expressly incorporated by reference herein.
The invention relates to a method and apparatus for detecting faults in components of a technical system. The apparatus uses fault-relevant process data whose condition, upon encountering a corresponding component fault, changes from a no fault condition to a fault condition when its condition value departs from a given tolerance.
Various types of fault diagnostic apparatus are known for the detection, identification and display of defective components of a production plant, a computer system, a motor vehicle, etc. Usually, the momentary actual condition values of the process variables of the system (which are composed of input values, output values and internal condition values) are detected and compared with given set values. If the momentary actual value differs from the set value by a given amount, the momentary value is evaluated and displayed as a fault. In electrical or electronic systems the evaluation can usually be performed directly by appropriate electronic means, for example, comparators, window discriminators, and the like. In some cases, in systems involving a mechanical factor, the corresponding process variables are converted by a measuring converter to an electrical signal which can be then evaluated by comparison.
One difficulty of such known apparatus is that the expression relating to the fault location or the nature of the fault is often ambiguous. Here, due to a lack of sensory equipment, the apparatus associates several possible component faults with a single fault signal. It is subsequently up to the operating personnel to make an evaluation of the fault display in order to find the correct and unequivocal one out of several or a number of possible fault signals. It is furthermore known to automatically determine the nature and location of a fault by a corresponding investment in sensing equipment and to then indicate the pertinent fault information. This fault information is in a coded or uncoded format and, if necessary, can be made usable for correction by operating or service personnel.
German Patent Document 41 24 542 C2 describes a fault diagnostic system for determining the cause of a fault in a apparatus under test, which system has a memory and a detecting system which detects the parameters of the apparatus under test. Stored in the memory system is a selection tree with nodes which correspond to the particular subunits of the apparatus under test, as well as test tables associated with the nodes, in which at least one parameter is supplied that is to be found by the detector system. Further stored in the memory system is a test condition relating thereto, plus a fault probability table corresponding to the results of tests according to at least one test condition, and names of slave nodes. In addition, at least two parameters to be detected and test conditions are given in a test table which is associated with a node having at least three slave nodes. Moreover, a search/interference system is stored beforehand in the memory system, and selects nodes along the selection tree and evaluates the corresponding test tables. Here, the nodes are chosen according to the result of the evaluation of the test tables. This is intended to achieve a targeted association of individual test tables by the search/interference system in the manner of a non-binary selection tree. The structure of the selection tree correspond to the hardware organization of the apparatus under test. This system requires relatively fast computations during the system's running time, since many decisions are to be made, and in some cases tables have to be reloaded.
U.S. Pat. No. 5,099,436 discloses a method and an apparatus for performing a system fault diagnosis which is based on a hybrid display of knowledge of the system to be diagnosed. Data obtained during the running time of the system are compared with an event-based system representation comprising a plurality of predefined events. An event is recognized when the data detected correlate to the critical parameters of the event. The recognized event and a corresponding set of ambiguous group S effects are analyzed, which characterize components which are to be re-sorted in an ambiguous group in accord with an associated sorting effect. Furthermore, a symptomatic fault model and a NOT function model can be analyzed in order to determine the symptomatic relationships and the nature of the NOT functions which are applicable to the running of the system. Each applicable symptom fault relationship and each kind of NOT function is associated with a set of ambiguity group effects which re-sorts the ambiguity group. Beginning with the component in the ambiguity group whose NOT function is the most probable, a structure model is analyzed. As a result of this analysis, proposals for repair are output with tests to be performed on the system.
This known procedure involves a current comprehensive data acquisition and constant comparison operations during system operation, and therefore a considerable amount of mathematical operations in the diagnostic section of the system. The system model describes the system components in an event structure with additional information on their probability of failure, ease of repair, accessibility, etc. The implementation of this diagnostic knowledge (for which special knowledge and/or experience are necessary) is not suitable for use where the systems to be diagnosed are subject to short-term changes in structure and character (as is the case in motor vehicles, for example).
Structural outlines for a computer-assisted fault diagnosis system for a motor vehicle are described in the following publication: N. Waleschkowski et al., Ein wissenbasiertes Fahrzeug-Diagnosesystem for den Einsatz in der Kfz-Werkstatt, Grundlagen und Anwendungen der känstlichen Intelligenz [A Knowledge-Based Diagnostic System for Use in the Automotive Industry, Bases and Applications of Artificial Intelligence] Zeitschrift künstliche Intelligenz KI 1/95, page 55. This system contains a diagnosis preparatory stage with a knowledge base which contains a structural model on the hierarchical construction of the technical system composed of individual secondary systems, and an effectiveness model on the effective relationships between the individual secondary systems. This system also contains a fault model which determines the course of the diagnosis and which shows the relationships between causes of faults and their effects as well as appropriate testing and repairs. A diagnosis performance stage interactively performs fault diagnoses by using the diagnosis program offered in the preparatory stage of the diagnosis.
An object of the present invention is to provide a method and fault diagnostic system (of the kind referred to above) with which system components which are suspected of faults can be recognized relatively fast, with comparatively few computing operations.
This and other objects and advantages are achieved by the method and fault diagnostic system according to the invention, in which upon the failure of a system component, i.e., upon the occurrence of a component fault, certain process factors known as “fault-relevant” factors will change their condition from a no-fault condition to a fault condition. As a result, it becomes possible from their condition to determine the one or more components suspected of being faulty. This binary condition decision for the particular process factor is performed according to whether the corresponding condition factor of the process variable lies within or outside of a preset tolerance range.
Furthermore, use can be made of the fact that knowledge of the operation of resources which are used not only by a fault signal path but also by one or more other signal paths can substantially reduce the number of the suspected components in the fault path.
For every component fault, the process va
Burkhardt Rainer
Strobel Herbert
Daimler-Chrysler AG
Evenson, McKeown, Edwards & Lenahan P.L.L.C.
Wright Norman M.
LandOfFree
Method and apparatus for detecting faults does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for detecting faults, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for detecting faults will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2465797