# US 2021/0081289 A1 · Method and system for assisting troubleshooting of a complex system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0081289 A1
- Document ID: US20210081289A1
- Status: Published application
- Kind: A1
- Application number: 16/956,850
- Filed: 2018-12-20
- Published: 2021-03-18
- Earliest claimed: 2017-12-22
- Assignees: Airbus Defence and Space GmbH; Airbus SAS
- Applicants: Airbus Defence and Space SAS; Airbus
- Inventors: Jean-Luc MARTY; David CANU; Alexandre ARNOLD
- Web page: https://patentgrep.com/patent/US20210081289A1

## CPC

- G06F 30/00: Computer-aided design [CAD]
- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06F 11/008: {Reliability or availability analysis}
- G06F 11/2257: {using expert systems}
- G06F 30/15: Vehicle, aircraft or watercraft design
- G06F 2111/02: CAD in a network environment, e.g. collaborative CAD or distributed simulation
- G06F 2111/08: Probabilistic or stochastic CAD
- G06F 2111/10: Numerical modelling
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system and a method for assisting with troubleshooting a complex system is disclosed in which the troubleshooting procedure can be modeled by a Markov decision process. Combining the fault tree technique with a Markov decision process, in order to determine in an optimal manner the sequence of troubleshooting actions will quickly address the consequences of a failure and ensure maintainability of the complex system.

## Claims

### Claim 1

A method for assisting with diagnosis of a complex system comprising a plurality of mechanical, electric, electronic and/or software components connected to each other, each among a portion of the components being associated with one or more fault sensors, the method, implemented by a processor, comprising:
receiving a file containing a modeling of some or all of the complex system by a fault tree defining some or all of the combinations of events causing a top event associated with a failure of the complex system, the fault tree comprising a plurality of intermediate events and a plurality of basic events,
assigning an a priori probability of occurrence value to each of the basic events of the fault tree, in order to generate an augmented fault tree,
assigning failure information to each of the basic events and each of the intermediate events of the augmented fault tree which are associated with a component under automatic monitoring, the failure information being obtained from fault sensors and describing an observation of the operating state of the component associated with the event of the augmented fault tree,
calculating an ex-post probability of occurrence value for each of the events of the augmented fault tree, based on the assigned a priori probability of occurrence values and the assigned failure information,
assigning a troubleshooting cost to each of the events of the augmented fault tree which are associated with a component not under automatic monitoring and serviceable, the troubleshooting cost defining the difficulty for the maintenance operator in identifying the failure of the component associated with the event of the augmented fault tree,
modeling, by a Markov decision process, MDP, based on the augmented fault tree, one or more possible evolutions of the complex system in response to at least one troubleshooting action by the maintenance operator on at least one component, the troubleshooting action corresponding to a procedure for identifying the failure or confirming the proper operation of the component,
determining an optimal sequence of troubleshooting actions for the complex system, according to a decision policy which minimizes the expected value of the sum of the troubleshooting costs and which is determined by applying an MDP-solving algorithm.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the step of calculating an ex-post probability of occurrence value comprises the steps of:
constructing a Bayesian network from the augmented fault tree, the Bayesian network being provided for making predictions concerning relationships between events of the augmented fault tree based on failure information and a priori probability values, and
calculating the probability value from the Bayesian network.

### Claim 3 (depends on 2)

The method according to claim 2, wherein the nodes of the Bayesian network represent corresponding components of the complex system, the arcs between nodes being associated with distributions of conditional probabilities which represent probabilities of events associated with certain components of the complex system which are associated with events associated with other components of the complex system.

### Claim 4 (depends on 1)

The method according to claim 1, wherein the step of assigning a troubleshooting cost comprises the step of:
obtaining the troubleshooting cost from a cost function dependent on at least one term which corresponds to a measurement or an estimate of the difficulty for the maintenance operator in identifying the failure of the component.

### Claim 5 (depends on 1)

The method according to claim 1, wherein the step of modeling by an MDP comprises the step of defining an MDP with a finite horizon described by a quadruplet {S,A,T,R} in which:
S is a set defining the states in which the complex system can be found at time t starting from an initial state of the complex system,
A is a set defining the troubleshooting actions that can be performed as a function of the state of the complex system and which influence the evolution of the current state of the complex system,
T is a transition function defining the set of probabilities of transitioning between two states of the complex system at states t and t+1, in response to the troubleshooting actions of set A,
R is a reward function defining the set of troubleshooting costs associated with the troubleshooting actions of set A.

### Claim 6 (depends on 5)

The method according to claim 5, wherein, when T defines the set of failure probabilities, a terminal state of the MDP is reached when at least one failure probability of a basic event is equal to 1.

### Claim 7 (depends on 5)

The method according to claim 5, wherein, when T defines the set of failure probabilities, a terminal state of the MDP is reached when the set of the failure probabilities is equal to 0.

### Claim 8 (depends on 5)

The method according to claim 5, wherein, when T defines the set of failure probabilities, a terminal state of the MDP is reached when all the serviceable nodes of the complex system have a failure probability equal to 0 or 1.

### Claim 9 (depends on 1)

A computer program product comprising instructions which, when executed by a processor, configure the processor to implement the steps of the method according to claim 1.

### Claim 10 (depends on 1)

A system for assisting with diagnosis of a complex system comprising a plurality of mechanical, electric, and/or software components connected to each other, each among a portion of the components being associated with one or more fault sensors, the system comprising a diagnosis assistance server and an electronic device both of which are arranged for accessing a communication network:
the diagnosis assistance server comprises a first processor configured to at least partly execute a method according to claim 1, based on a file containing a modeling of some or all of the complex system by a fault tree defining some or all of the combinations of events causing a top event associated with a failure of the complex system, the fault tree comprising a plurality of intermediate events and a plurality of basic events,
the portable electronic device comprises a second processor coupled to a user input device and to a display device, in which the second processor is configured for:
receiving all or part of the optimal sequence of troubleshooting actions for the complex system, via the communication network,
displaying on the display device an image representative of the current troubleshooting action to be performed, based on the optimal sequence of troubleshooting actions for the complex system,
detecting an input signal from the user input device, which is representative of a user input during the displaying of the image representative of the current troubleshooting action to be performed,
in response to a user input which is indicative of detection of a failure or confirmation of the proper operation of the component associated with the current troubleshooting action to be performed, displaying on the display device an image representative of the next troubleshooting action to be performed, based on the optimal sequence of troubleshooting actions for the complex system.

### Claim 11 (depends on 10)

The system according to claim 10, wherein:
the second processor of the portable electronic device is further configured for sending the user input to the first processor of the diagnostic assistance server,
in response to receiving the user input, the first processor is further configured for:
determining the next troubleshooting action from the optimal sequence of troubleshooting actions for the complex system, and
sending the next troubleshooting action, from the optimal sequence of troubleshooting actions for the complex system to the second processor of the portable electronic device.

### Claim 12 (depends on 10)

The system according to claim 10, wherein the user input device is comprised in the display device.
