Automatic fault detection and diagnosis in complex physical systems
OwnerPaul Rensing, Keith Corbett +1inventors
InventorsPaul Rensing, Keith Corbett, Charles Frankston
Life of this patent
Estimating the term…
Abstract
A fault detection and diagnosis (FDD) system designed to automatically and efficiently determine the true fault underlying a set of symptoms, presenting the operator with a clear cause and effect diagram for all faults and symptoms. An Event Cluster is used to group related symptoms and causes, potentially across multiple pieces of equipment, and establish a cause and effect relationship chain.
Claims14
A method for automated detection of ultimate faults in a multiple component control system, the method comprising:
- a) collecting data from each sensor among a plurality of sensors, each of the plurality of sensors corresponding to a respective one of the multiple components within the control system;
- b) performing a separate set of time-series calculations on the data collected from each of the sensors;
- c) converting results of each set of time-series calculations into a plurality of events;
- d) applying first generated sets of cluster rules stored in a database of a fault detection system, each first generated set of cluster rules being associated with similar sets of events to link together the similar sets of events into respective event ranges and to link together the respective event ranges to form an event cluster at a common fault, and a second generated set of cluster rules stored in the database, the second generated set of cluster rules being associated with the formed event cluster and at least one similar event to link together the event cluster and the at least one similar event into respective second event ranges and to link the respective second event ranges to form an event cluster at a common fault:
- and
- e) applying additional generated sets of cluster rules associated with a respective formed event cluster and at least one similar event to link together the respective event clusters and the at least one similar event into respective additional event ranges and to link the additional event ranges to form an event cluster at a new common fault until an ultimate fault diagnosis is determined.
A fault detection system for a multiple component control system, the fault detection system comprising:
- a) a plurality of sensors, each sensor corresponding to at least one of the plurality of components to detect a status of the corresponding component;
- b) a fault detection and diagnostics device coupled to each of the plurality of sensors, the fault detection and diagnostics device comprising:
- a memory including at least one database configured to store a plurality of cluster rules in the form of pattern analysis templates;
- one of more processors being configured:
- to collect the status data from each of the sensors of the corresponding component, the data collected from each sensor being performed in integrals specific to the corresponding component, and to perform a separate set of time-series calculations for the data collected from each sensor;
- to convert the results of each set of time-series calculations into a plurality of events;
- to apply first generated sets of cluster rules, each first generated set of cluster rules being associated with similar sets of events to link together the similar sets of events into respective event ranges and to link together the respective event ranges to form an event cluster at a common fault, and a second generated set of cluster rules, the second set of cluster rules being associated with the formed event cluster and at least one similar event to link together the event cluster and the at least one similar event into respective second event ranges and to link the respective second event ranges to form n event cluster at a common fault, and
- to apply additional generated sets of cluster rules associated with a respective formed event cluster and at least one similar event to link together respective event clusters and the at least one similar event into respective additional event ranges and to link the additional event ranges to form an event cluster at a new common fault until a final fault diagnosis is determined.
Dependencies are inferred from the claim text.