Method of identifying abnormal behavior in a fleet of vehicles

Data processing: measuring – calibrating – or testing – Measurement system – Remote supervisory monitoring

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C702S185000, C701S029000

Reexamination Certificate

active

06732063

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to generally to the field of data mining, and in particular to of monitoring a complex system having multiple sub-systems described by a plurality of operating parameters. The invention is particularly well suited to identifying abnormal behavior in a fleet of vehicles, but it can be applied to other complex systems. The invention is primarily applicable to aircraft, but could be applied to other types of vehicle. Such behavior might, for example, relate to engine operation or other critical flight systems. The behavior is measured in parameters that might include exhaust gas temperature, core vibration, fuel flow etc.
2. Description of Related Art
Efficient operation and maintenance of modern aircraft, particularly “fly-by-wire” aircraft involves processing large amounts of sensor data. Sensors continually monitor all aspects of flight systems. Proper use of the data requires an understanding of its contents, the ability to analyze it properly, and the ability to identify when and in which part of a large fleet of aircraft a particular condition or performance parameter is not within the allowable range. Proper alerts must be generated for maintenance staff along with an explanation of the probable cause for an abnormal situation.
This requires access to a variety of data, the most important part of which is sensor measurement/data included in individual aircraft reports (e.g. engine cruise report, engine, take off report). These reports are generated on a regular basis pursuant to a request or when a deviation of one or more parameters occur. The operation and proper maintenance of aircraft becomes more difficult as the airline fleet becomes larger and more and more parameters need to be closely monitored. Engineers and fleet specialists must constantly be aware of the status of all aircraft systems and must be able to investigate any abnormal behaviors as soon as they occur. The investigation requires an explanation as what may be the cause of any abnormal situation.
One possible approach to the above problem is to analyze aircraft parametric data, using a data analysis tool, and search for patterns or useful information in the data that may explain the problem. Almost all commercial data analysis tools assume that the data collected from a particular process (such as the operation of an aircraft) can be easily analyzed, trends and patterns accurately recognized and any discoveries presented to the user in an understandable way. Aerospace is one of the domains in which these assumptions are not valid. Even when a data analysis tool is selected, all users must have sufficient training and time to acquire the data, properly use the data analysis tool and go through the ordinary process of data analysis in order to discover a useful knowledge that they need. However, in the aerospace industry, like many other industries, it is difficult to expect engineers and technicians to follow the normal data mining path. Several reasons exist for this. All the required data may not be integrated into one database management system. The engineers and operators do not have sufficient time to analyze huge amounts of data, unless there is an urgent requirement. Complexity of the data analysis process is beyond the ordinary tools that they have access to, in most cases. There is no well defined automated mechanism to extract, pre-process and analyze the data, and summarize the results so that the engineers and technicians can use it. Even when there are tools available for data analysis, these tools are specific for certain tasks defined by the vendors.
There are two possible approaches to this problem. One approach is the batch mode in which an aerospace engineer has to select a given amount of historical data and having a problem in mind uses a commercial data mining software tool to analyze the data and search for useful patterns in the data. This is usually an iterative and time consuming process that may help in identifying an abnormal situation in the operation of the aircraft for which the data was collected.
In the batch mode, the data collection and selection process is time consuming and has to be accurate. The user has to have a good understanding of the data mining process and the problem for which the data mining tool is used. In many cases, there is a need for data cleaning, as there may be various problems in the data, such as out-of-range data, missing data, etc. The user needs to have some leads as to where there is a problem so that relevant data is selected and analyzed. In many cases, even when there are some results, it may be too late for aerospace engineers to properly use them. Finally, the user needs an explanation as to what may be the cause of any problem.
The other approach is the on-line mode, in which a vendor (such as an engine manufacturer) provides a software application that can be used to generate alerts about the performance of certain system parameters for which the software has been designed. The problem with this mode is that the airline staff need to acquire dedicated software for every system for which they have the data and want to monitor performance. This is not possible for all systems on board today's aircraft as there is no monitoring software for every system for which the data is available. Even when such software systems are used, airline staff can only monitor parameters that these software systems are capable of monitoring. An example is SAGE software provided by General Electric to monitor certain parameters of aircraft engines. Systems such as SAGE generate alerts that are not necessarily relevant for a given airline, due to different modes of operation. Finally, there is no explanation to support any identified abnormal situation as to what maybe the cause of that situation.
A data mining system has been described in the above SAGE system that continually gathers data from a number of sources and generates alerts when certain normalized thresholds are exceeded. While this is an excellent potential technique for solving the above problems, it does not take into account the realities of the real world, where one needs to take into account the performance of the fleet as a whole. For example, as the fleet as a whole ages, the nominal thresholds will shift.
There is a need to a monitoring system that will alleviate these shortcomings of the prior art.
SUMMARY OF THE INVENTION
According to the present invention there is provided a method of monitoring a complex system having multiple sub-systems described by a plurality of operating parameters, comprising continually generating data pertaining to said operating parameters during operation of said system; storing said data in a central database; defining a window of samples over which data is to be analyzed; normalizing said data to take into account variability factors and introduce a weighting factor to define thresholds dependent on the performance of individual monitored components and the performance of said components across the fleet; storing said defined thresholds for said defined window; and continually comparing new incoming data from said sub-systems with said stored defined thresholds to identify abnormalities in the system.
The weighting factor in the calculation of the threshold is important to allow the effects of the sub-systems, typically aircraft, and the complex system, typically the fleet, to be balanced. The weighting factor prevents the generation of unnecessary alerts. Introduction of the weighting factor resulted from extensive experiments in which a weighting factor was not used and the effects of both aircraft and fleet mean and standard deviation were equally taken into account.
The invention permits all aircraft systems (e.g. auxiliary power units, main engines) for which data is available to be continuously monitored using an integrated software system. The monitoring is performed so that the fleet specialists and engineers are informed of conditions when there is a deviation in the

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Method of identifying abnormal behavior in a fleet of vehicles 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 of identifying abnormal behavior in a fleet of vehicles, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of identifying abnormal behavior in a fleet of vehicles will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3219846

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.