Automatic repair prediction based on sensor data

Abstract

To provide a system and a method for automated prediction of repair based on sensor data.SOLUTION: In a system for repairing a machine, an apparatus includes a computer-readable program code for a knowledge-based system including an inference engine and a knowledge base, and processing circuitry configured to access memory and execute the code. The apparatus is configured to: at least receive a time series of measurements of operating conditions of the machine recorded during an operation of the machine; cluster the time series of measurements into one or more corresponding clusters; identify a pattern across the clusters; define a current state of the machine that includes the pattern across the clusters; search a knowledge base for a past case describing a solution corresponding to a past abnormal state similar to the current state, where the corresponding solution includes a repair action; and generate an output display indicating the repair action to address the current state.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001]

[0001] The present disclosure relates generally to machine diagnostics and repair, and more particularly to automatically predicting repairs based on sensor data. [Background technology]

[0002]

[0002] Machines, including vehicles such as aircraft, spacecraft, ships, automobiles, and trains, typically include some type of performance monitoring system that records data regarding the machine's performance, including the performance of the machine's various systems (and subsystems). This data includes a record of specific performance events that occur during the machine's operation. The performance monitoring system usually collects the data and reports all collected data to a user. The user can then use the data to determine what type of maintenance or repair, if any, the machine may require. For example, if the data indicates that a particular mechanical or electromechanical system of the machine is malfunctioning or that the performance of one or more mechanical or electromechanical systems may contribute to a future machine failure, the user can perform appropriate repairs to the machine at the next opportunity.

[0003]

[0003] Current systems for machine performance and failure monitoring provide users with the data necessary to make appropriate repair decisions, but still require the user to sort through all of the data to determine the most appropriate repair action to address the failure mode. Therefore, the user must sort through and interpret the data in light of the user's knowledge of the particular machine. This can be time-consuming, and the first repair action performed is not necessarily the most appropriate repair action, especially for complex machines such as aircraft and other vehicles. For many types of machines, especially commercial vehicles, the amount of time the vehicle is out of service is expensive for the vehicle owner. Therefore, the longer it takes for the most appropriate repair action to be performed for a given failure mode, the longer the vehicle will be out of service. This can be costly for the vehicle owner when the vehicle could otherwise be in service.

[0004]

[0004] It would therefore be desirable to have a system and method that takes into account at least some of the above-mentioned issues, as well as other possible issues. Summary of the Invention

[0005]

[0005] Exemplary embodiments of the present disclosure are directed to machine diagnosis and repair, and in particular to the automated prediction of repairs for machines such as aircraft based on sensor data. The exemplary embodiments recommend correct repair actions using historical data, operational state measurements, and repair actions performed to address the current state of the machine. The exemplary embodiments use a knowledge-based system to search a knowledge base for repair actions from the historical data. The knowledge-based system identifies matching or similar past cases from the current state of the machine and current and past operational states, and searches for repair actions for the current state of the machine from the past repair actions.

[0006]

[0006] Exemplary embodiments provide more informed repair of a machine's mechanical or electromechanical systems, enabling the machine to perform fast and accurate repair actions for failure modes in those systems. In the case of aircraft, this can reduce the time the aircraft is out of service and increase on-schedule performance for airlines. In the defense sector, exemplary embodiments can result in increased aircraft availability and improved mission readiness.

[0007]

[0007] Accordingly, the present disclosure includes, but is not limited to, the following exemplary embodiments.

[0008]

[0008] Some exemplary embodiments provide a method including receiving time series measurements of a plurality of operational states of a machine recorded during operation of the machine; clustering the time series measurements into one or more respective groups (clusters) of measurements; identifying patterns across the plurality of groups of measurements; defining a current state of the machine comprising the patterns across the plurality of groups of measurements; accessing a knowledge base including a set of past cases describing a plurality of past abnormal states of the machine and respective solutions, wherein the plurality of past abnormal states of the machine include a plurality of patterns across the plurality of groups of measurements of a plurality of operational states of the machine recorded during a plurality of previous instances of machine operation during which a plurality of failure modes of a plurality of systems of the machine occurred, and wherein each solution includes a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base for past cases describing respective solutions to past abnormal states of a plurality of past abnormal states that are similar to the current state, wherein each solution is identified as a candidate solution for the current state, and wherein each solution to the past abnormal state includes a repair action; and generating an output display showing the repair actions to address the current state.

[0009]

[0009] In some exemplary embodiments of the methods of the above exemplary embodiments or any combination of the above exemplary embodiments, the time series measurements are clustered based on input parameters including a fixed number of groups, a maximum distance between a central measurement and other measurements in a group of measurements, or a minimum number of measurements in a group of measurements.

[0010]

[0010] In some exemplary embodiments of the methods of the above exemplary embodiments or any combination of the above exemplary embodiments, time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for noisy applications (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm.

[0011]

[0011] In some exemplary embodiments of the methods of the aforementioned exemplary embodiments or any combination of the aforementioned exemplary embodiments, the machine is a transport, the plurality of systems includes a plurality of transport systems, and receiving the time series measurements further includes receiving fault data from a transport system of the plurality of transport systems that was recorded during an operation that was or included movement of the transport, and the method further includes diagnosing from the fault data a current condition that is a failure mode of the transport system or another transport system of the plurality of transport systems.

[0012]

[0012] In some exemplary embodiments of the methods of the aforementioned exemplary embodiments or any combination of the aforementioned exemplary embodiments, diagnosing the current condition includes diagnosing a timing of the failure mode, and the method further includes identifying a plurality of operating states associated with the failure mode, where the plurality of operating states is a subset of a larger plurality of operating states of the transporter, identifying a time interval from the timing of the failure mode, and extracting a time series measurement for the time interval from a larger plurality of time series measurements of the larger plurality of operating states.

[0013]

[0013] In some exemplary embodiments of the methods of the aforementioned exemplary embodiments or any combination of the aforementioned exemplary embodiments, the plurality of repair actions in the knowledge base are weighted based on the success rate of the repair action for addressing each failure mode of the plurality of failure modes, and one or more of a plurality of past abnormal conditions of one or more of the plurality of past cases matches the current condition or is within a specified margin of match, and searching the knowledge base includes selecting the most highly weighted repair action, thereby the past case having the highest success rate among the plurality of repair actions of the one or more of the plurality of past cases.

[0014]

[0014] In some exemplary embodiments of the methods of the aforementioned exemplary embodiments or any combination of the aforementioned exemplary embodiments, the machine is an aircraft and the current state is indicated by a failure mode reported by the aircraft during an operation that is or includes the flight of the aircraft.

[0015]

[0015] Some exemplary embodiments provide an apparatus comprising: a memory configured to store computer-readable program code for a knowledge-based system including an inference engine and a knowledge base; and a processing circuit configured to access the memory and execute the computer-readable program code so as to cause the inference engine and thereby the apparatus to perform at least a method of any of the aforementioned exemplary embodiments or any combination of any of the exemplary embodiments.

[0016] Some exemplary embodiments provide a computer-readable storage medium having stored thereon computer-readable program code and a knowledge-based system, the knowledge-based system including an inference engine and a knowledge base, the computer-readable program code, upon execution by a processing circuit, causing the inference engine to cause an apparatus to perform at least the method of any preceding exemplary embodiment or any combination of any of the exemplary embodiments.

[0017]

[0017] The above and other features, aspects, and advantages of the present disclosure will become apparent from the following detailed description taken in conjunction with the accompanying drawings, which are briefly described below. In specific exemplary embodiments described herein, the present disclosure includes any combination of two, three, four, or more features or elements specified in this disclosure, whether such features or elements are explicitly combined or otherwise listed. This disclosure is intended to be read holistically, whereby any separable features or elements of the present disclosure, in any aspect and exemplary embodiment thereof, should be considered as combinable unless otherwise specified in the context of the disclosure.

[0018]

[0018] Therefore, it should be understood that this Summary is provided solely for the purpose of summarizing some exemplary implementations to provide a basic understanding of some aspects of the present disclosure. As such, it should be recognized that the exemplary embodiments described above are examples only and should not be construed in any way to narrow the scope or nature of the present disclosure. Other exemplary implementations, aspects, and advantages will become apparent from the following detailed description taken in conjunction with the accompanying drawings, which illustrate, by way of example, the principles of some exemplary implementations described.

[0019]

[0019] Having described exemplary embodiments of the present disclosure in general terms above, reference is now made to the accompanying drawings, which are not necessarily drawn to scale, and in which: [Brief explanation of the drawings]

[0020] [Figure 1] 1 illustrates an aircraft according to some exemplary embodiments of the present disclosure. [Figure 2]

[0021] 1 illustrates a system for diagnosing and repairing a mechanical or electromechanical system of a machine, such as an aircraft, according to an exemplary embodiment of the present disclosure. [Figure 3A]

[0022] 10 shows an example of a confusion matrix for sample results of a training method according to an exemplary embodiment. [Figure 3B] 10 shows an example of a confusion matrix for sample results of a training method according to an exemplary embodiment. [Figure 3C] 10 shows an example of a confusion matrix for sample results of a training method according to an exemplary embodiment. [Figure 3D] 10 shows an example of a confusion matrix for sample results of a training method according to an exemplary embodiment. [Figure 4]

[0023] 1 is a flowchart illustrating various steps in a method for diagnosing and repairing a mechanical or electromechanical system of a machine, such as an aircraft, according to an exemplary embodiment of the present disclosure. [Figure 5]

[0024] 1 illustrates an apparatus according to some exemplary embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0021]

[0025] Certain embodiments of the present disclosure are now more comprehensively described below with reference to the accompanying drawings. The accompanying drawings illustrate some, but not all, embodiments of the present disclosure. Indeed, various implementations of the present disclosure may be embodied in many different forms and should not be construed as limited to the implementations set forth herein. Rather, these exemplary implementations are provided so that this disclosure will be comprehensive and complete, and will fully convey the scope of the disclosure to those skilled in the art. For example, unless otherwise specified, referring to something as a "first," "second," etc. should not be construed as implying a particular order. Also, something that may be described as being above something else (unless otherwise specified) may instead be below, and vice versa. Similarly, something that may be described as being to the left of something else may instead be to the right, and vice versa. Like reference numerals refer to like elements throughout.

[0022]

[0026] Exemplary embodiments of the present disclosure relate generally to machine repair, and more particularly to repairing mechanical or electromechanical systems of machines. While the exemplary embodiments are described primarily in conjunction with repairing mechanical or electromechanical systems of aircraft, it should be understood that the exemplary embodiments may be utilized in conjunction with a variety of other machines. Examples of suitable machines that may benefit from the exemplary embodiments include vehicles such as spacecraft, watercraft, automobiles, and trains. Furthermore, it should be understood that, unless otherwise specified, the terms "data," "content," "digital content," "information," and similar terms may sometimes be used interchangeably. Similarly, the terms "maintenance" and "repair" may sometimes be used interchangeably.

[0023]

[0027] Exemplary embodiments of the present disclosure provide for more informed repair of mechanical or electromechanical systems of machines, such as vehicles. This can result in the most appropriate initial repair for each system's failure mode. Some exemplary embodiments assume that each failure type is often due to one or more root causes from an exhaustive list of root causes. For each specific root cause of a failure mode, the operating conditions of the machine may exhibit a significant degree of similarity. Thus, the root cause(s) of each failure occurrence may be reflected in the operating conditions of the machine (recorded during machine operation), including more specifically the operating conditions leading up to and immediately prior to the occurrence of the failure mode.

[0024]

[0028] FIG. 1 illustrates one type of aircraft 100 that may benefit from exemplary embodiments of the present disclosure. As illustrated, the aircraft includes a fuselage 102 having a fuselage 104, wings 106, and a tail section 108. The aircraft also includes a high-level system 110, such as a propulsion system. In the particular example illustrated in FIG. 1, the propulsion system includes two wing-mounted engines 112. In other examples, the propulsion system may include other configurations, such as engines carried by other portions of the aircraft, including the fuselage and / or tail section. The high-level system may also include an electrical system 114, a hydraulic system 116, and / or an environmental system 118. Any number of other systems may also be included.

[0025]

[0029] FIG. 2 illustrates a system 200 for repairing a machine, such as a vehicle (e.g., aircraft 100), according to an exemplary embodiment of the present disclosure. The repair may include repairing a mechanical or electromechanical system of the machine. The system may include any of several different subsystems (each individual system) for performing one or more functions or operations. As shown, in some embodiments, the system includes at least one source 202 of data. In some embodiments, the source includes memory that may be located in a single source or distributed across multiple sources. The data may be stored in several different ways, such as in a database or file in any of several different types or formats (e.g., Quick Access Recorder (QAR), Enhanced Airborne Flight Recorder (EAER), Flight Data Recorder (FDR), etc.).

[0026]

[0030] In some embodiments where the machine is an aircraft, the data includes indications of failure modes of systems of the aircraft, some of which may be indicated by flight deck effects (FDE) reported by the aircraft. One example of a suitable failure mode is a failure of the aircraft's cabin air compressor (CAC). The data also includes flight data about flights of the aircraft, including previous flights of the aircraft. The flight data includes time-series measurements of multiple operating conditions of the aircraft recorded during flight from sensors or avionics systems onboard the aircraft. Examples of suitable operating conditions include flow rate, temperature, pressure, recorded values ​​from the CAC, altitude, airspeed, etc.

[0027]

[0031] In some more particular examples, the flight data includes time-series measurements of multiple operating conditions recorded during a flight by a monitoring system from sensors or avionics systems, and at least some of the flight data for each flight is accessible from reports generated by the monitoring system. The monitoring system may be onboard a machine, such as aircraft 204, which in some examples corresponds to aircraft 100. In this regard, the aircraft may be configured to record the flight data and generate reports, which may be transmitted wirelessly, directly, via satellite 206, or via a network, to a particular source 202 of data. Source 202 is illustrated as, and sometimes referred to as, an aircraft health management (AHM) system 202a. In still other examples, the flight data may be transmitted via a wired connection or a portable data storage device (e.g., flash memory, thumb drive).

[0028]

[0032] System 200, in an exemplary embodiment of the present disclosure, includes a knowledge-based system 208 for determining, from at least some of the above-described data, repair actions to address failure modes of mechanical or electromechanical systems of a machine (e.g., aircraft 204). The knowledge-based system includes an inference engine 210 and a knowledge base 212. Multiple subsystems, including source(s) 202, AHM system 202a, and the knowledge base system including the inference engine and knowledge base, may be co-located or directly coupled to one another, or, in some embodiments, various ones of the multiple subsystems may communicate with one another across one or more computer networks 214. Furthermore, while illustrated as part of system 200, any one or more of the above may function or operate as a separate system without involvement of any of the other subsystems. It should also be understood that the system may include one or more additional or alternative subsystems other than those shown in FIG. 2 .

[0029]

[0033] According to some example embodiments of the present disclosure, inference engine 210 is configured to receive a plurality of time-series measurements of aircraft operating conditions recorded during flight from sensors or avionics systems onboard the aircraft. The time-series measurements include operational data, including normal data and fault data. The inference engine is configured to cluster the time-series measurements into one or more respective groups of measurements, and the inference engine is configured to identify patterns 218 across the groups of measurements and define a current state 220 of the machine comprising the patterns across the groups of measurements. In some examples, depending on the clustering algorithm(s) utilized by the inference engine, the time-series measurements are clustered based on input parameters, including a fixed number of groups, a maximum distance between a central measurement and other measurements in a group of measurements, or a minimum number of measurements in a group of measurements.

[0030]

[0034] In some embodiments, inference engine 210 is configured to access knowledge base 212, which includes a set of past cases 222 describing past abnormal conditions 224 and respective solutions 226. The past abnormal conditions include patterns 230 across multiple groups of measurements of machine operating conditions recorded during previous instances of machine operation during which a failure mode of the machine's systems occurred (e.g., for each of previous flights by an aircraft's onboard sensors or an aircraft's onboard flight recorder from the avionics system). Each solution includes multiple repair actions 232 that were performed to address each of the multiple failure modes. In some embodiments, the multiple repair actions are extracted from maintenance records, such as from an avionics resource management system. An example of an appropriate repair action for a CAC failure may include instructions to inspect and / or repair the CMSC Left 2 CAC, the CMSC Right 2 CAC, or the Nitrogen Generating System Motor Driven Compressor (NGS MDC).

[0031]

[0035] Inference engine 210, in some embodiments, is configured to search knowledge base 212 for past cases 222 that describe respective solutions (among respective solutions 226) for past abnormal conditions (which may be the past abnormal condition most similar to the current condition) among multiple past abnormal conditions 224 that are similar to current condition 220. In this regard, each solution is identified as a candidate solution for the current condition. In some embodiments, this search includes searching knowledge base 212 for past cases that describe past abnormal conditions that include a pattern among multiple patterns 230 that is most similar to pattern 218 for the current condition. Each solution for a past abnormal condition that is identified as a candidate solution for the current condition includes repair action 234 (among multiple repair actions 232). The inference engine is then configured to generate output display 236 that displays repair actions to address the machine (e.g., aircraft 204) and / or its mechanical or electromechanical systems. In some embodiments, the repair actions refer to instructions 238 for performing the repair actions to address the current condition.

[0032]

[0036] In some embodiments, knowledge base system 208 is constructed in several different ways. In this regard, past abnormal conditions 224 of a machine's (e.g., aircraft 204) systems that occurred during previous flights of the aircraft may be collected along with time-series measurements of the aircraft's operational states recorded during each of the previous flights, and repair actions 232 performed to address each of the failure modes. This data may be collected for multiple systems over a period of one year or more for multiple previous flights for which information on repair actions performed is available. Data analytics, natural language processing (NLP), machine learning, etc. may be performed on the data to develop inference engine 210 to identify patterns 230 across multiple groups of measurements of the operational states, which may then be stored in knowledge base 212.

[0033]

[0037] In some embodiments, the plurality of repair actions 232 in knowledge base 212 are weighted based on the success rate of the plurality of repair actions for addressing each of the plurality of failure modes. When one or more of the plurality of past abnormal conditions 224 of one or more of the plurality of past cases 222 matches or is within a specified margin of match with the current condition 220, inference engine 210 may select the most highly weighted repair action, thereby the past case having the highest success rate among the plurality of repair actions of the one or more of the plurality of past cases.

[0034]

[0038] In some examples, system 200 further includes a reasoner, such as reasoner 240, configured to receive failure data from a system (e.g., an aircraft system recorded during operation of aircraft 204), the failure data being a portion of operational data that includes failure data and normal data. In these examples, the reasoner is configured to diagnose, from the failure data, a current state 220 of the mechanical or electromechanical system that is a failure mode of the aircraft system or another aircraft system of the plurality of aircraft systems. The reasoner is configured to operate according to a clustering algorithm or a suite of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering of applications with noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm. As shown in FIG. 2 , the reasoner is positioned onboard the aircraft and operates in real time or near real time on the failure data. In other examples, the reasoner may not be onboard the aircraft and may be configured to communicate with other of the subsystems across computer network(s) 214, or the reasoner may be co-located with knowledge-based system 208. In these examples, operational, normal, or fault data may be downloaded from the aircraft during or after the flight.

[0035]

[0039] In some further embodiments, reasoner 240 is configured to diagnose, from the failure data, a current condition that is or is indicated by a failure mode of the aircraft system (mechanical or electromechanical system) or another aircraft system among the plurality of aircraft systems. The failure mode is reported by aircraft 204 during operations that are or include a flight of the aircraft. The diagnosing may be according to a clustering algorithm or a set of clustering algorithms developed based on machine learning techniques. In some of these further embodiments, diagnosing the current condition includes diagnosing a timing of the failure mode, and the reasoner identifies a plurality of operating conditions as associated with the failure mode. The plurality of operating conditions may be a subset of a larger plurality of operating conditions of aircraft 204. Also, in some of these further embodiments, the reasoner is configured to identify a time interval from the timing of the failure mode and extract a time series measurement for the time interval from a larger plurality of time series measurements of the operating conditions.

[0036]

[0040] As described above, exemplary embodiments may be utilized in conjunction with a variety of machines, including vehicles such as spacecraft, ships, automobiles, trains, etc. More generally, then, inference engine 210 is configured to receive a plurality of time-series measurements of the machine's operating state recorded during operation. The inference engine is configured to cluster the time-series measurements into one or more respective groups of measurements, identify patterns 218 across the groups of measurements, and define a current state 220 comprising the patterns across the groups of measurements.

[0037]

[0041] The inference engine 210 is configured to access a knowledge base 212 that includes a set of past cases 222 that describe past abnormal conditions 224 and respective solutions 226. The past abnormal conditions include patterns 230 across groups of measurements of operational conditions of the machine that were recorded during previous instances of machine operation during which failure modes of systems of the machine occurred. Each solution includes repair actions 232 that were performed to address a respective one of the failure modes.

[0038]

[0042] The inference engine 210 is configured to search the knowledge base 212 for past cases that describe solutions to past abnormal conditions (which may be the past abnormal condition that is most similar to the current condition) among a plurality of past abnormal conditions 224 that are similar to the current condition 220. Each solution is identified as a candidate solution for the current condition, and each solution to the past abnormal condition includes a repair action 234. The inference engine is then configured to generate an output display 236 that displays the repair action to address the current condition. In some embodiments, the repair action references instructions 238 for performing the repair action to address the current condition.

[0039]

[0043] A training method for the inference engine 210 may include various steps, beginning with inputting sensor data as time-series measurements. The measurements may be compiled into a matrix. If the number of sensor data points is too large, the data may be compressed using a clustering algorithm or a set of clustering algorithms. Examples of clustering algorithms include the k-means algorithm, the k-medoids algorithm (when some of the sensor variables are binary or discrete), DPGMM, and / or DBSCAN. The output of the clustering algorithm(s) is each represented by the mean value of the points within a grouping, where the centers may be referred to as "states."

[0040]

[0044] Each sensor data point may be assigned to a state (which may also be referred to as "labeling") by calculating the distance between the sensor data point and each state (center of the cluster). The sensor data point is then assigned to the state with the shortest calculated distance. After the sensor data point is assigned to a state, a sequence of latency states for compressed data or a sequence of uncompressed raw data may be calculated within a sliding window of n seconds. The size of the sliding window (n) may vary depending on the use of raw or compressed data. In some embodiments, n varies from 5 to 60 seconds, although other values ​​of n less than 5 or greater than 60 may also be appropriately utilized.

[0041]

[0045] Either the raw data or the sequence of latency states may be compared to data including multiple past cases 222. In some embodiments, this data and / or past cases may include data including summaries of past repair actions related to failure modes and stored in Positive Repair Identification Text Analyzer (PRITA) software. The PRITA data may be compared to the calculated sequence to associate repair actions with features present in the condensed or raw data sequence.

[0042]

[0046] The training method may incorporate a supervised learning algorithm such as random forest, although other machine learning algorithms (e.g., neural networks) may also be utilized. Once trained, the inference engine 210 may receive input (data) in a window size of n and output repair actions based on and related to the input.

[0043]

[0047] FIGS. 3A, 3B, 3C, and 3D show example confusion matrices for sample results of a training method using random forests, according to some example implementations. Rows in the matrix correspond to true repairs, and columns correspond to repairs predicted by an inference engine (e.g., inference engine 210). Each of FIGS. 3A-3D uses an example fault message associated with a CAC, which may also be referred to as a common motor start controller (CMSC). FIGS. 3A and 3B show respective sample results calculated for raw data and latency conditions associated with the fault message "CMSC right 2 CAC detected internal or feeder fault." The example of FIG. 3A yields approximately 95% recall and approximately 98% precision, while the example of FIG. 3B yields approximately 89% recall and 97-98% precision. FIGS. 3C and 3D show respective sample results calculated for raw data and latency conditions associated with the fault message "CMSC left 2 CAC has no output to any bus." The example of FIG. 3C yields approximately 100% recall and approximately 99% precision, while the example of FIG. 3D yields approximately 97-98% recall and approximately 95% precision. While the resulting recall and precision from raw data versus latency are similar in these examples, the amount of data (i.e., compressed data) sent downstream using latency is significantly lower than the amount of data sent downstream using raw data (e.g., over 1,900 kilobytes of raw data per failure message versus approximately 60 bytes of compressed data per failure message). The latency examples (FIGS. 3B and 3D) may span longer durations. Consequently, the sliding window size n in these examples is larger than the n for the raw data examples of FIG. 3A and FIG. 3C.

[0044]

[0048] FIG. 4 is a flowchart illustrating various steps in a method 400. According to an exemplary embodiment of the present disclosure, the method includes executing a knowledge base system 208, which includes an inference engine 210 and a knowledge base 212, where the inference engine may perform various steps of the method. As illustrated at block 402, the method includes receiving a time series of measurements of a plurality of operational conditions of the machine recorded during operation of the machine. In some embodiments, receiving the time series of measurements further includes receiving fault data from a vehicle system of the plurality of vehicle systems recorded during an operation that is or includes movement of the vehicle. As illustrated at block 404, the time series of measurements are clustered into one or more respective groups of measurements. As illustrated at blocks 406 and 408, patterns 218 are identified across the groups of measurements, and a current state of the machine is defined that includes the patterns across the groups of measurements.

[0045]

[0049] As illustrated at block 410, the inference engine 210 accesses a knowledge base including a set of past cases 222 that describe multiple past abnormal conditions 224 of the machine and respective solutions 226. The multiple past abnormal conditions of the machine include patterns 230 across multiple groups of measurements of multiple operational conditions of the machine recorded during multiple previous instances of machine operation during which a failure mode of the machine's systems occurred. Each solution includes repair actions 232 that were performed to address a respective one of the multiple failure modes.

[0046]

[0050] As depicted at block 412, the inference engine 210 searches the knowledge base 212 for past cases that describe solutions to each past abnormal condition among the plurality of past abnormal conditions 224 that are similar to the current condition 220 (which may be the past abnormal condition that is most similar to the current condition). In some embodiments, searching the knowledge base includes selecting the past case that has the highest weighted repair action, thereby having the highest success rate among one or more of the plurality of repair actions among the plurality of past cases in the set of past cases 222. Each solution to the plurality of past abnormal conditions that has been identified as a candidate solution for the current condition includes a repair action. As depicted at block 414, the inference engine then generates an output display 236 that displays the repair action to address the current condition.

[0047]

[0051] In some examples, the method 300 further includes diagnosing a current condition 220 that is a failure mode of the transport system or another transport system of the plurality of transport systems from the failure data. Diagnosing the current condition may include diagnosing a timing of the failure mode. The method may also include identifying a plurality of operating conditions associated with the failure mode, identifying a time interval from the timing of the failure mode, and extracting a time series of measurements for the time interval from the more time series of measurements of the more plurality of operating conditions.

[0048]

[0052] According to exemplary embodiments of the present disclosure, system 200 and its subsystems, including source 202, AHM system 202a, and knowledge base system 208, including inference engine 210 and knowledge base 212, may be implemented by various means. Means for implementing the system and its subsystems may include hardware, alone or under instructions of one or more computer programs from a computer-readable storage medium. In some examples, one or more devices may be configured to function as or otherwise implement the systems and their subsystems illustrated and described herein. In examples including two or more devices, the respective devices may be connected to or otherwise able to communicate with each other in several different ways (e.g., directly or indirectly, such as via a wired or wireless network).

[0049]

[0053] 5 illustrates an apparatus 500 according to some exemplary embodiments of the present disclosure. Generally, an apparatus according to exemplary embodiments of the present disclosure may comprise, include, or be embodied in one or more fixed or portable electronic devices. Examples of suitable electronic devices include smartphones, tablet computers, laptop computers, desktop computers, workstation computers, server computers, etc. The apparatus may include one or more of several components, such as, for example, a processing circuit 502 (e.g., a processor unit) coupled to a memory 504 (e.g., a storage device).

[0050]

[0054] The processing circuit 502 may be comprised of one or more processors, either alone or in combination with one or more memories. A processing circuit is generally any computer hardware capable of processing information, such as, for example, data, computer programs, and / or other suitable electronic information. A processing circuit may be comprised of a collection of electronic circuits, some of which may be packaged as an integrated circuit or multiple interconnected integrated circuits (integrated circuits may more commonly be referred to as a "chip"). The processing circuit may be configured to execute a computer program, which may be stored on-board the processing circuit or otherwise stored in memory 504 (of the same or another device).

[0051]

[0055] The processing circuitry 502 may be several processors, a multiprocessor core, or some other type of processor, depending on the particular implementation. Furthermore, the processing circuitry may be implemented using several heterogeneous processor systems (where a main processor resides on a single chip with one or more secondary processors). As another example, the processing circuitry may be a symmetric multiprocessor system that includes multiple homogeneous processors. In yet another example, the processing circuitry may be embodied as or otherwise include one or more ASICs, FPGAs, or the like. Thus, while the processing circuitry may be capable of executing a computer program to perform one or more functions, various examples of the processing circuitry may be capable of performing one or more functions without the aid of a computer program. In any case, the processing circuitry may be appropriately programmed to perform functions or operations in accordance with exemplary embodiments of the present disclosure.

[0052]

[0056] The memory 504 is generally any computer hardware capable of temporarily and / or permanently storing information, such as, for example, data, computer programs (e.g., computer-readable program code 506), and / or other suitable information. The memory may include volatile and / or non-volatile memory and may be fixed or removable. Examples of suitable memory include random access memory (RAM), read-only memory (ROM), hard drives, flash memory, thumb drives, removable computer diskettes, optical disks, magnetic tape, or any combination thereof. Optical disks may include compact disks (read-only memory (CD-ROM), read / write (CD-R / W)), DVDs, etc. In various instances, the memory may be referred to as a computer-readable storage medium. A computer-readable storage medium is a non-transitory device capable of storing information and may be distinguished from a computer-readable transmission medium (such as a transient electronic signal capable of conveying information from one place to another). Computer-readable media as described herein may generally refer to computer-readable storage media or computer-readable transmission media.

[0053]

[0057] In addition to the memory 504, the processing circuit 502 may also be connected to one or more interfaces for displaying, transmitting, and / or receiving information. The interfaces may include a communication interface 508 (e.g., a communication unit) and / or one or more user interfaces. The communication interface may be configured to transmit and / or receive information, for example, from other device(s), network(s), etc. The communication interface may be configured to transmit and / or receive information over physical (wired) and / or wireless communication links. Examples of suitable communication interfaces include a network interface controller (NIC), a wireless NIC (WNIC), etc.

[0054]

[0058] The user interface may include a display 510 and / or one or more user input interfaces 512 (e.g., input / output units). The display may be configured to present or otherwise display information to a user, suitable examples of which include a liquid crystal display (LCD), a light-emitting diode display (LED), a plasma display panel (PDP), etc. The user input interface may be wired or wireless and may be configured to receive information from a user, for example, for processing, storage, and / or display within the device. Suitable examples of user input interfaces include a microphone, an image or video capture device, a keyboard or keypad, a joystick, a touch-sensitive surface (separate from or integrated into a touchscreen), a biometric sensor, etc. The user interface may further include one or more interfaces for communicating with peripheral devices (e.g., a printer, a scanner, etc.).

[0055]

[0059] As indicated above, program code instructions may be stored in a memory and executed by a processing circuit (programmed with the program code instructions) to implement the functions of the systems, subsystems, tools, and corresponding elements described herein. It will be appreciated that any suitable program code instructions may be loaded from a computer-readable storage medium into a computer or other programmable device to cause a particular machine to become a means for implementing the functions identified herein. Such program code instructions may be stored on a computer-readable storage medium, which may instruct a computer, processing circuit, or other programmable device to function in a particular manner to produce a particular machine or a particular article of manufacture. The instructions stored on a computer-readable storage medium may produce an article of manufacture, which becomes a means for performing the functions described herein. The program code instructions may be read from a computer-readable storage medium and loaded into a computer, processing circuit, or other programmable device to configure the computer, processing circuit, or other programmable device to perform steps to be performed on or by the computer, processing circuit, or other programmable device.

[0056]

[0060] The reading, loading, and execution of the program code instructions may be performed serially, such that one instruction is read, loaded, and executed at a time. In some exemplary implementations, the reading, loading, and / or execution may be performed in parallel, such that multiple instructions are read, loaded, and / or executed together. Execution of the program code instructions may produce a computer-implemented process, such that the instructions, when executed by a computer, processing circuit, or other programmable device, provide steps for implementing the functions described herein.

[0057]

[0061] Execution of the instructions by a processing circuit or storage of the instructions in a computer-readable storage medium supports a combination of steps to perform a particular function. Thus, device 500 may include a processing circuit 502 and a computer-readable storage medium or memory 504 coupled to the processing circuit, the processing circuit configured to execute computer-readable program code 506 stored in the memory. It will also be understood that one or more functions and combinations of functions may be performed by a computer system based on special purpose hardware and / or a processing circuit that performs particular functions, or a combination of special purpose hardware and program code instructions.

[0058]

[0062] Numerous modifications and other embodiments of the disclosure set forth herein will come to mind to one skilled in the art to which this disclosure pertains having the benefit of the teachings presented in the foregoing description and the associated drawings. It is to be understood, therefore, that the disclosure is not limited to the particular implementations disclosed, and that variations and other implementations are intended to be within the scope of the appended claims. Furthermore, while the foregoing description and the associated drawings describe exemplary implementations in light of particular example combinations of elements and / or functions, it should be recognized that alternative implementations may provide various combinations of elements and / or functions without departing from the scope of the appended claims. That is, other combinations of elements and / or functions than those expressly described above are contemplated, as may be set forth in some of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

[0059]

[0063] Clause 1 1. An apparatus (500) comprising: a memory (504) configured to store computer-readable program code (506) for a knowledge-based system (208); and a processing circuit (502) configured to access the memory and execute the computer-readable program code to cause an inference engine (210) to perform operations, the operations including at least receiving a time series of measurements of a plurality of operational states of the machine recorded during operation of the machine; clustering the time series of measurements into one or more respective groups of measurements; identifying a pattern (218) across the plurality of groups of measurements; defining a current state (220) of the machine comprising the pattern across the plurality of groups of measurements; accessing a knowledge base (212) including a set of past cases (222) describing a plurality of past abnormal states (224) of the machine and respective solutions (226); an output display (506) displaying the repair actions (232) to address the current condition; and generating an output display (506) displaying the repair actions (234) to address the current condition.

[0060]

[0064] Clause 2 The apparatus (500) described in clause 1, wherein the time series measurements are clustered based on input parameters including a fixed number of groups, a maximum distance between a central measurement and other measurements in a group of measurements, or a minimum number of measurements in a group of measurements.

[0061]

[0065] Clause 3 3. The apparatus (500) of clause 1 or 2, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for applications with noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm.

[0062]

[0066] Clause 4 The apparatus (500) of any one of clauses 1 to 3, wherein the machine is a vehicle, the plurality of systems includes a plurality of vehicle systems, and the apparatus executing receiving the time series of measurements further includes the apparatus executing receiving, from a vehicle system of the plurality of vehicle systems, fault data recorded during the operation, which is or includes movement of the vehicle, and the processing circuit (502) further executes the computer readable program code (506) to cause the inference engine (210) thereby to cause the apparatus to diagnose, from the fault data, the current state (220), which is a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

[0063]

[0067] Clause 5 The apparatus (500) of clause 4, wherein the apparatus performing diagnosing the current state (220) includes the apparatus performing diagnosing the timing of the failure mode, and the apparatus further performs identifying the plurality of operating states as associated with the failure mode, the plurality of operating states being a subset of a larger plurality of operating states of the vehicle, identifying a time interval from the timing of the failure mode, and extracting the time series measurement value for the time interval from a larger plurality of time series measurement values ​​of the larger plurality of operating states.

[0064]

[0068] Clause 6 6. The apparatus of claim 1, wherein a plurality of repair actions in the knowledge base are weighted based on a success rate of the plurality of repair actions for addressing the respective failure modes of the plurality of failure modes, and wherein one or more of the plurality of past abnormal conditions of one or more of the plurality of past cases matches or is within a specified margin of match with the current condition, and wherein searching the knowledge base includes selecting the most highly weighted repair action, thereby the past case having the highest success rate among the one or more repair actions of the plurality of past cases.

[0065]

[0069] Clause 7 The apparatus (500) of any one of clauses 1 to 6, wherein the machine is an aircraft (204) and the current state (220) is indicated by a failure mode reported by the aircraft during the operation, which is or includes a flight of the aircraft.

[0066]

[0070] Article 8 receiving (402) a time series of measurements of a plurality of operational states of the machine recorded during operation of the machine; clustering (404) the time series of measurements into one or more respective groups of measurements; identifying (406) a pattern (218) across the plurality of groups of measurements; defining (408) a current state (220) of the machine comprising the pattern across the plurality of groups of measurements; and accessing (410) a knowledge base (212) comprising a set of past cases (222) describing a plurality of past abnormal states (224) of the machine and respective solutions (226), wherein the plurality of past abnormal states of the machine are a combination of measurements of the plurality of operational states of the machine recorded during a plurality of previous instances of operation of the machine during which a plurality of failure modes of a plurality of systems of the machine occurred; a plurality of patterns (230) spanning a number of groups, each solution including a plurality of repair actions (232) performed to address a respective failure mode among the plurality of failure modes; searching the knowledge base (212) for past cases describing respective solutions to past abnormal conditions among the plurality of past abnormal conditions similar to the current condition, each solution being identified as a candidate solution for the current condition, each solution for the past abnormal condition including a repair action; and generating an output display (236) displaying the repair actions (234) to address the current condition.

[0067]

[0071] Article 9 9. The method (400) of clause 8, wherein the time series measurements are clustered based on input parameters including a fixed number of groups, a maximum distance between a central measurement and other measurements in a group of measurements, or a minimum number of measurements in a group of measurements.

[0068]

[0072] Article 10 10. The method (400) of clause 8 or 9, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for applications with noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Models (DPGMM) algorithm.

[0069]

[0073] Article 11 11. The method (400) of any one of clauses 8 to 10, wherein the machine is a vehicle, the plurality of systems includes a plurality of vehicle systems, and receiving the time series of measurements further includes receiving fault data from a vehicle system of the plurality of vehicle systems that was recorded during the operation that is or includes movement of the vehicle, and the method further includes diagnosing from the fault data the current condition (220) being a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

[0070]

[0074] Article 12 12. The method of claim 11, wherein diagnosing the current state includes diagnosing a timing of the failure mode, the method further including identifying the plurality of operating states as associated with the failure mode, the plurality of operating states being a subset of a larger plurality of operating states of the vehicle; identifying a time interval from the timing of the failure mode; and extracting the time series of measurements for the time interval from a larger plurality of time series of measurements of the larger plurality of operating states.

[0071]

[0075] Article 13 13. The method (400) of any one of clauses 8 to 12, wherein a plurality of repair actions (232) in the knowledge base (212) are weighted based on a success rate of the plurality of repair actions for addressing the respective failure mode of the plurality of failure modes, and wherein one or more of the plurality of past abnormal conditions (224) of one or more of the plurality of past cases (222) match or are within a specified margin of match with the current condition (220), and wherein searching the knowledge base includes selecting the most highly weighted repair action, thereby the past case having the highest success rate among the one or more repair actions of the plurality of past cases.

[0072]

[0076] Article 14 14. The method (400) of any one of clauses 8 to 13, wherein the machine is an aircraft (204) and the current state (220) is indicated by a failure mode reported by the aircraft during the operation, which is or includes a flight of the aircraft.

[0073]

[0077] Article 15 a computer-readable storage medium (504) having stored thereon computer-readable program code (506) and a knowledge-based system (208), the knowledge-based system including an inference engine (210) and a knowledge base (212), the computer-readable program code, when executed by a processing circuit (502), causing the inference engine to perform at least the following steps: receive a time series of measurements of a plurality of operational states of the machine recorded during operation of the machine; cluster the time series of measurements into one or more respective groups of measurements; identify a pattern (218) across the plurality of groups of measurements; define a current state (220) of the machine comprising the pattern across the plurality of groups of measurements; access a knowledge base (212) including a set of past cases (222) describing a plurality of past abnormal states (224) of the machine and respective solutions (226); a computer-readable storage medium (504) that causes the computer to: access a knowledge base (212), the normal state including a plurality of patterns (230) across a plurality of groups of measurements of the plurality of operational states of the machine recorded during a plurality of previous instances of operation of the machine during which a plurality of failure modes of a plurality of systems of the machine occurred, the respective solutions including a plurality of repair actions (232) that were performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base for past instances that describe respective solutions to past abnormal states of the plurality of past abnormal states that are similar to the current state, the respective solutions being identified as candidate solutions for the current state, the respective solutions for the past abnormal states including repair actions; and generating an output display (236) that displays the repair actions (234) to address the current state.

[0074]

[0078] Article 16 16. The computer-readable storage medium (504) of clause 15, wherein the time series measurements are clustered based on input parameters including a fixed number of groups, a maximum distance between a central measurement and other measurements in a group of measurements, or a minimum number of measurements in a group of measurements.

[0075]

[0079] Article 17 17. The computer-readable storage medium (504) of clause 15 or 16, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a Density-Based Spatial Clustering for Applications with Noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm.

[0076]

[0080] Article 18 18. The computer-readable storage medium (504) of any one of clauses 15 to 17, wherein the machine is a vehicle, the plurality of systems includes a plurality of vehicle systems, and wherein the device (500) performing the receiving of the time series measurements further comprises the device performing the receiving of fault data from a vehicle system of the plurality of vehicle systems that was recorded during the operation that is or includes movement of the vehicle, and the device further performs the diagnosing, from the fault data, the current state (220) being a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

[0077]

[0081] Article 19 19. The computer-readable storage medium (504) of clause 18, wherein the device performing diagnosing the current state (220) includes the device performing diagnosing the timing of the failure mode, and the device (500) further performs identifying the plurality of operating states as associated with the failure mode, the plurality of operating states being a subset of a larger plurality of operating states of the vehicle, identifying a time interval from the timing of the failure mode, and extracting the time series of measurements for the time interval from a larger plurality of time series of measurements of the larger plurality of operating states.

[0078]

[0082] Article 20 20. The computer-readable storage medium (504) of any one of clauses 15 to 19, wherein a plurality of repair actions (232) in the knowledge base (212) are weighted based on a success rate of the plurality of repair actions for addressing the respective failure modes of the plurality of failure modes, and one or more of the plurality of past abnormal conditions (224) of one or more of the plurality of past cases (222) match or are within a specified margin of match with the current condition (220), and wherein the device (500) performing searching the knowledge base includes the device performing selecting the most highly weighted repair action, thereby the past case having the highest success rate among the one or more repair actions of the plurality of past cases.

[0079]

[0083] Article 21 21. The computer-readable storage medium (504) of any one of clauses 15 to 20, wherein the machine is an aircraft (204) and the current state (220) is indicated by a failure mode reported by the aircraft during the operation, which is or includes a flight of the aircraft.

Claims

1. An apparatus for automating aircraft repair forecasting, comprising: a communications device onboard the aircraft configured to communicate with aircraft sensor devices that monitor the aircraft's electrical, hydraulic, or aircraft propulsion operating system; a memory device onboard the aircraft configured to store computer readable program code for a knowledge base system including an inference engine and a knowledge base; and a processing circuit onboard the aircraft configured to access the memory device and execute the computer readable program code to cause the inference engine, and thereby the apparatus, to perform operations, the operations including: receiving, via the communication device, from the aircraft sensor device monitoring an operating system of the aircraft, a plurality of time-series measurements of real-time operational conditions of the aircraft recorded during flight operation of the aircraft; clustering the time series measurements into respective groups of measurements; and identifying a pattern across the plurality of groups of measurements, wherein identifying the pattern comprises: compressing the sensor data points of the time series of measurements to form compressed data, wherein the compressing includes: calculating a distance between a data point of the sensor data point and a cluster state representing the mean value of the data points in one of the groups of measurements; assigning the cluster state to the data point based on determining that the calculated distance between the data point and the cluster state is the shortest distance among a set of distances calculated between the data point and a plurality of cluster states of the plurality of groups; compressing the sensor data points to form compressed data based on sequencing latency states for the compressed data calculated over a sliding window of n seconds; identifying patterns across the groups of measurements, including: defining a current state of the aircraft including the pattern across a plurality of groups of the measurements; accessing a knowledge base including a set of past cases describing a plurality of past abnormal conditions of the aircraft and respective solutions, the plurality of past abnormal conditions of the aircraft including a plurality of patterns across a plurality of groups of measurements of the plurality of operational conditions of the aircraft recorded during a plurality of previous instances of aircraft operation during which a plurality of failure modes of a plurality of systems of the aircraft occurred, and the respective solutions including a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base to identify past cases that describe respective solutions to past abnormal conditions among the plurality of past abnormal conditions that are similar to the current condition, the respective solutions being identified as candidate solutions for the current condition, and the respective solutions to the past abnormal conditions including repair actions; and generating an output display that displays the repair action to address the current condition.

2. An apparatus, a memory configured to store computer readable program code for a knowledge-based system including an inference engine and a knowledge base; a processing circuit configured to access the memory and execute the computer readable program code to cause the inference engine, and thereby the device, to perform operations, the operations including at least: receiving a time series of measurements of a plurality of operating conditions of the machine recorded during operation of the machine; clustering the time series of measurements into one or more respective groups of measurements; identifying patterns across groups of said measurements; defining a current state of the machine including the pattern across multiple groups of the measurements; accessing a knowledge base including a set of past cases describing a plurality of past abnormal conditions of the machine and respective solutions, the plurality of past abnormal conditions of the machine including a plurality of patterns across a plurality of groups of measurements of the plurality of operational conditions of the machine recorded during a plurality of previous instances of operation of the machine during which a plurality of failure modes of a plurality of systems of the machine occurred, and the respective solutions including a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base for past cases that describe respective solutions to past abnormal conditions of the plurality of past abnormal conditions similar to the current condition, the respective solutions being identified as candidate solutions for the current condition, and the respective solutions to the past abnormal conditions including repair actions; a plurality of repair actions in the knowledge base are weighted based on a success rate of the plurality of repair actions for addressing the respective failure mode of the plurality of failure modes; wherein one or more of the plurality of past abnormal conditions of one or more of the plurality of past cases matches the current condition or is within a defined margin of match, and searching the knowledge base selects the most highly weighted repair action, thereby the past case having the highest success rate among the plurality of repair actions of the one or more of the plurality of past cases; and generating an output display that displays the repair action to address the current condition.

3. 3. The apparatus of claim 1, wherein the clustering of the time series measurements is further based on input parameters including a fixed number of groups.

4. 3. The apparatus of claim 1 or 2, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for noisy applications (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm.

5. the plurality of systems includes a plurality of vehicle systems, and the time series of measurements includes failure data from a vehicle system of the plurality of vehicle systems recorded during the operation that is or includes movement of the aircraft; 10. The apparatus of claim 1, wherein the processing circuitry further causes the apparatus to diagnose from the failure data the current condition being a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

6. The method of claim 6, wherein determining the current state of the aircraft includes the device diagnosing the timing of the failure mode, and the device: identifying the plurality of operating conditions as associated with the failure mode, the plurality of operating conditions being a subset of a larger plurality of operating conditions of the vehicle; Identifying a time interval from the timing of the failure mode; and The apparatus of claim 5 , further comprising: extracting the time series of measurements for the time interval from a greater number of time series of measurements of the greater number of operating conditions.

7. The device described in claim 1, wherein the current state is indicated by a failure mode reported by the aircraft during the operation which is or includes a flight of the aircraft.

8. A method for automating forecasting of aircraft repairs, comprising: receiving, via a communication device onboard the aircraft, time-series measurements of a plurality of operational conditions of the aircraft recorded during flight operation of the aircraft from aircraft sensor devices monitoring the aircraft's electrical, hydraulic, or aircraft propulsion operating system; clustering the time series of measurements into respective groups of measurements via processing circuitry on board the aircraft; and identifying, via the processing circuitry, a pattern across the plurality of groups of measurements, wherein identifying the pattern comprises: compressing the sensor data points of the time series of measurements to form compressed data, wherein the compressing includes: calculating a distance between a data point of the sensor data point and a cluster state representing the mean value of the data points in one of the groups of measurements; assigning the cluster state to the data point based on determining that the calculated distance between the data point and the cluster state is the shortest distance among a set of distances calculated between the data point and a plurality of cluster states of the plurality of groups; compressing the sensor data points to form compressed data based on sequencing latency states of said compressed data calculated over a sliding window of n seconds; identifying patterns across the groups of measurements, including: determining, via the processing circuitry, a current state of the aircraft including the pattern across a plurality of groups of the measurements; accessing, via the processing circuitry, a knowledge base including a set of past cases describing a plurality of past abnormal conditions of the aircraft and respective solutions, the plurality of past abnormal conditions of the aircraft including a plurality of patterns across a plurality of groups of measurements of the plurality of operational conditions of the aircraft recorded during a plurality of previous instances of aircraft operation during which a plurality of failure modes of a plurality of systems of the aircraft occurred, and the respective solutions including a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; identifying, via the processing circuitry searching the knowledge base, past cases that describe respective solutions to past abnormal conditions among the plurality of past abnormal conditions that are similar to the current condition, the respective solutions being identified as candidate solutions for the current condition, and the respective solutions to the past abnormal conditions including repair actions; and generating an output display that indicates the repair action to address the current condition.

9. The method of claim 8 , wherein clustering the time series measurements is further based on input parameters including a fixed number of groups.

10. 9. The method of claim 8, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for applications with noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Model (DPGMM) algorithm.

11. The plurality of systems includes a plurality of vehicle systems, and the time series measurements include failure data from vehicle systems of the plurality of vehicle systems recorded during the flight operation that is or includes movement of the aircraft; 10. The method of claim 8, further comprising diagnosing the current condition from the failure data as a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

12. The method of claim 1, wherein determining the current state of the aircraft includes diagnosing the timing of the failure mode. identifying the plurality of operating conditions as associated with the failure mode, the plurality of operating conditions being a subset of a larger plurality of operating conditions of the vehicle; Identifying a time interval from the timing of the failure mode; and The method of claim 11 , comprising extracting the time series of measurements for the time interval from a greater number of time series of measurements of the greater number of operating conditions.

13. a plurality of repair actions in the knowledge base are weighted based on a success rate of the plurality of repair actions for addressing the respective failure mode of the plurality of failure modes; 9. The method of claim 8, wherein one or more of the plurality of past abnormal conditions of one or more of the plurality of past cases matches the current condition or is within a specified margin of match, and searching the knowledge base includes selecting the most highly weighted repair action, thereby the past case having the highest success rate among the plurality of repair actions of the one or more of the plurality of past cases.

14. The method of claim 8, wherein the current state is indicated by a failure mode reported by the aircraft during the flight operation that is or includes a flight of the aircraft.

15. A computer-readable storage medium having stored therein computer-readable program code and a knowledge-based system, the knowledge-based system including an inference engine and a knowledge base, the computer-readable program code being executed by a processing circuit to cause the inference engine, and thereby an aircraft-mounted device, to: receiving, via a communication device onboard the aircraft, time-series measurements of a plurality of operational conditions of the aircraft recorded during flight operation of the aircraft from aircraft sensor devices monitoring the aircraft's electrical, hydraulic, or propulsion aircraft operating system; clustering the time series measurements into respective groups of measurements; and identifying a pattern across the plurality of groups of measurements, wherein identifying the pattern comprises: compressing the sensor data points of the time series of measurements to form compressed data, wherein the compressing includes: calculating a distance between a data point of the sensor data point and a cluster state representing the mean value of the data points in one of the groups of measurements; assigning the cluster state to the data point based on determining that the calculated distance between the data point and the cluster state is the shortest distance among a set of distances calculated between the data point and a plurality of cluster states of the plurality of groups; compressing the sensor data points to form compressed data based on sequencing latency states of said compressed data calculated over a sliding window of n seconds; identifying patterns across the plurality of groups of measurements, including defining a current state of the aircraft including the pattern across a plurality of groups of the measurements; accessing a knowledge base including a set of past cases describing a plurality of past abnormal conditions of the aircraft and respective solutions, the plurality of past abnormal conditions of the aircraft including a plurality of patterns across a plurality of groups of measurements of the plurality of operational conditions of the aircraft recorded during a plurality of previous instances of aircraft operation during which a plurality of failure modes of a plurality of systems of the aircraft occurred, and the respective solutions including a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base to identify past cases that describe respective solutions to past abnormal conditions among the plurality of past abnormal conditions that are similar to the current condition, the respective solutions being identified as candidate solutions for the current condition, and the respective solutions to the past abnormal conditions including repair actions; and generating an output display that indicates the repair actions to address the current condition.

16. A computer-readable storage medium having stored therein computer-readable program code and a knowledge-based system, the knowledge-based system including an inference engine and a knowledge base, the computer-readable program code being executed by a processing circuit to cause the inference engine, and thereby the device, to perform at least: receiving a time series of measurements of a plurality of operating conditions of the machine recorded during operation of the machine; clustering the time series of measurements into one or more respective groups of measurements; identifying patterns across groups of said measurements; defining a current state of the machine including the pattern across multiple groups of the measurements; accessing a knowledge base including a set of past cases describing a plurality of past abnormal conditions of the machine and respective solutions, the plurality of past abnormal conditions of the machine including a plurality of patterns across a plurality of groups of measurements of the plurality of operational conditions of the machine recorded during a plurality of previous instances of operation of the machine during which a plurality of failure modes of a plurality of systems of the machine occurred, and the respective solutions including a plurality of repair actions performed to address a respective failure mode of the plurality of failure modes; searching the knowledge base for past cases that describe respective solutions to past abnormal conditions of the plurality of past abnormal conditions similar to the current condition, the respective solutions being identified as candidate solutions for the current condition, and the respective solutions to the past abnormal conditions including repair actions; a plurality of repair actions in the knowledge base are weighted based on a success rate of the plurality of repair actions for addressing the respective failure mode of the plurality of failure modes; wherein one or more of the plurality of past abnormal conditions of one or more of the plurality of past cases matches the current condition or is within a defined margin of match, and wherein the device performs searching the knowledge base; and wherein the device performs selecting the past case having the highest weighted repair action, thereby having the highest success rate among the plurality of repair actions of the one or more of the plurality of past cases; generating an output display that indicates the repair action to address the current condition.

17. 17. The computer-readable storage medium of claim 15 or 16, wherein clustering the time series measurements is further based on input parameters including a fixed number of groups.

18. 17. The computer-readable storage medium of claim 15 or 16, wherein the time series measurements are clustered into a fixed number of groups according to a clustering algorithm or a set of clustering algorithms, including a k-means clustering algorithm, a density-based spatial clustering for applications with noise (DBSCAN) algorithm, or a Dirichlet Process Gaussian Mixture Models (DPGMM) algorithm.

19. The plurality of systems includes a plurality of vehicle systems, and the time series measurements include failure data from vehicle systems of the plurality of vehicle systems recorded during the operation that is or includes movement of the aircraft; 16. The computer-readable storage medium of claim 15, wherein the device is further configured to diagnose, from the failure data, the current condition as a failure mode of the vehicle system or another vehicle system of the plurality of vehicle systems.

20. The method of claim 20, wherein determining the current state of the aircraft includes diagnosing the timing of the failure mode, the method comprising: identifying the plurality of operating conditions as associated with the failure mode, the plurality of operating conditions being a subset of a larger plurality of operating conditions of the vehicle; Identifying a time interval from the timing of the failure mode; and 20. The computer-readable storage medium of claim 19, further comprising extracting the time series of measurements for the time interval from a greater number of time series of measurements of the greater number of operating conditions.

Images