Prediction and detection of abnormalities in aircraft equipment

Abstract

To provide methods and systems for detecting an operational anomaly before a fault condition occurs.SOLUTION: A method includes obtaining sensor data captured by a sensor of an aircraft during a power up event. The sensor data includes multiple parameter values, each corresponding to a sample period. The method further includes determining a set of delta values, each indicating a difference between parameter values for consecutive sample periods of the sensor data. The method further includes determining a set of quantized delta values by assigning the delta values to quantization bins based on magnitudes of the delta values. The method further includes determining a normalized count of delta values for each quantization bin. The method further includes comparing the normalized counts of delta values to anomaly detection thresholds. The method further includes generating, based on the comparisons, output indicating whether the sensor data is indicative of an operational anomaly.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates generally to anomaly prediction and detection in aviation equipment. [Background technology]

[0002] Modern aircraft are highly complex, with many integrated systems that interact to provide features such as increased efficiency, improved safety, and improved performance. A failure of any of these systems can require significant expenditure of time and resources to correct the failure and prepare the aircraft for future operations. Additionally, performing unscheduled maintenance to correct the failure condition can reduce aircraft availability. For example, depending on the location of the aircraft when the failure condition occurs, the parts, equipment, or technicians required to perform the unscheduled maintenance may not be available. In either case, the parts, equipment, or technicians may need to be transported to the aircraft, resulting in significant delays. Summary of the Invention [Problem to be solved by the invention]

[0003] Aircraft operators plan many maintenance activities in an attempt to avoid this situation. Planned maintenance is performed according to a plan with the objective of, at least in part, preventing failures from occurring. While planned maintenance may avoid many failure conditions, it is inefficient because it results in the replacement of parts that are assumed to be in good working order (e.g., have never failed). [Means for solving the problem]

[0004] In certain embodiments, the aircraft includes a sensor that captures sensor data during a power-up event. The sensor data includes a plurality of parameter values, each parameter value corresponding to a respective sample period during the power-up event. The aircraft further includes one or more processors and a memory device that stores the sensor data and instructions executable by the one or more processors to detect an operational anomaly of an aircraft device associated with the sensor. When the instructions are executed by the one or more processors, the one or more processors perform operations including determining a set of delta values. Each delta value in the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods of the sensor data. The operations further include determining a set of quantized delta values ​​by assigning a delta value from the set of delta values ​​to a quantization bin based on a magnitude of the delta value. The operations further include determining a normalized coefficient of the delta value for each quantization bin. The operations further include performing a comparison between the normalized coefficient of the delta value for a particular quantization bin and an anomaly detection threshold for the particular quantization bin. The operations further include generating an output indicative of whether the sensor data is indicative of an operational anomaly of the aviation equipment based on the comparison.

[0005] In another particular embodiment, a method includes acquiring, with a computing device, sensor data captured by a sensor of the aircraft during a power-up event. The sensor data includes a plurality of parameter values, each parameter value corresponding to a respective sample period during the power-up event. The method further includes determining, with the computing device, a set of delta values. Each delta value in the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods of the sensor data. The method further includes determining, with the computing device, a set of quantized delta values ​​by assigning a delta value from the set of delta values ​​to a quantization bin based on a magnitude of the delta value. The method further includes determining, with the computing device, a normalized count of the delta values ​​for each quantization bin. The method further includes performing, with the computing device, a comparison between the normalized count of the delta values ​​for a particular quantization bin and an anomaly detection threshold for the particular quantization bin. The method further includes generating, by the computing device and based on the comparison, an output indicative of whether the sensor data is indicative of an operational anomaly of the aviation equipment associated with the sensor.

[0006] In another particular embodiment, a non-transitory computer readable medium stores instructions that, when executed by a processor, cause the processor to initiate, perform, or control operations. The operations include obtaining sensor data captured by a sensor of the aircraft during a power-up event. The sensor data includes a plurality of parameter values, each parameter value corresponding to a respective sample period during the power-up event. The operations further include determining a set of delta values. Each delta value in the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods of the sensor data. The operations further include determining a set of quantized delta values ​​by assigning a delta value from the set of delta values ​​to a quantization bin based on a magnitude of the delta value. The operations further include determining a normalized coefficient of the delta values ​​for each quantization bin. The operations further include performing a comparison between the normalized coefficient of the delta values ​​for a particular quantization bin and an anomaly detection threshold for the particular quantization bin. The operations further include generating an output indicative of whether the sensor data is indicative of an operational anomaly of an aviation equipment associated with the sensor based on the comparison.

[0007] In another particular embodiment, a method of training an anomaly detection system includes obtaining, with a computing device, a plurality of sets of sensor data captured by on-board sensors of one or more aircraft during a plurality of power-up events. Each set of sensor data corresponds to a respective power-up event of a respective aircraft and includes a plurality of parameter values. Each parameter value corresponds to a respective sample period during a respective power-up event. The method further includes determining, with the computing device, a set of delta values ​​for each set of sensor data. Each delta value from the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods from the set of sensor data. The method further includes determining, with the computing device, a set of quantized delta values ​​by assigning delta values ​​of the set of delta values ​​to quantization bins based on magnitudes of the delta values. The method further includes determining, with the computing device, a normalized coefficient of the delta values ​​for each quantization bin. The method further includes setting an anomaly detection threshold for each quantization bin based on the normalized coefficient of the delta values ​​and fault indication data for the one or more aircraft.

[0008] The features, functions, and advantages described herein can be achieved separately in various embodiments or may be combined in yet further embodiments, further details of which can be seen with reference to the following description and drawings. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 illustrates an example of a system that includes one or more aircraft and an anomaly detection system. [Diagram 2] 2 is a diagram showing details of operations performed by the anomaly detection system of FIG. 1 to detect anomalies based on sensor data. [Diagram 3] FIG. 2 is a diagram showing details of operations performed to train the anomaly detection system of FIG. 1. [Figure 4]2 is a flowchart of an example method for detecting anomalies using the anomaly detection system of FIG. 1. [Diagram 5] 2 is a flowchart of an example method for determining an anomaly detection threshold of the anomaly detection system of FIG. 1. [Figure 6] FIG. 1 is a block diagram of a computing environment including computing devices configured to support aspects of the computer-implemented methods and computer-executable program instructions (or codes) according to the present disclosure. [Figure 7] 2 is a flowchart of an example of a life cycle of an aircraft equipped with the anomaly detection system of FIG. 1. [Figure 8] FIG. 2 is a block diagram of an aircraft equipped with the anomaly detection system of FIG. 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] The disclosed system and method allows for detection of abnormal operation of aircraft equipment before a fault condition actually occurs. By detecting anomalies before a fault condition occurs, aircraft operators can plan maintenance more effectively and efficiently. For example, rather than replacing parts based on flight hours (which can be wasteful if the part does not actually need to be replaced), parts can be replaced before they fail when an anomaly detection system indicates that the part will soon need replacing. As a result, inefficiencies from replacing parts that do not need replacing and inefficiencies resulting from the aircraft being in a fault condition are both avoided or reduced.

[0011] The drawings and the following description show certain exemplary embodiments. It will be understood that those skilled in the art can devise various configurations that embody the principles described herein and fall within the scope of the claims that follow this description, even if not explicitly described or shown herein. Furthermore, any examples described herein are intended to be helpful in understanding the principles of the present disclosure and should not be construed as limiting. As a result, the present disclosure is not limited to the specific embodiments or examples described below, but rather by the claims and their equivalents.

[0012] Certain embodiments are described herein with reference to the drawings. In the description, common features are indicated by common reference numbers throughout the drawings. In some drawings, multiple instances of a particular type of features are used. Although these features are physically and / or logically different, the same reference number is used for each, and the different instances are distinguished by adding a letter to the reference number. When referring to a group or type of features (e.g., when no particular one of the features is referenced), the reference number without a distinguishing letter is used. However, when referring to one particular feature of multiple features of the same type, the reference number with a distinguishing letter is used. For example, with reference to FIG. 3, multiple examples of aircraft are shown and are associated with reference numbers 102A and 102K. When referring to a particular one of these aircraft, a distinguishing letter is used. By way of example, when referring to a first aircraft 102A, the distinguishing letter "A" is used. However, when referring to any arbitrary aircraft, or to the aircraft as a group, the reference number 102 without a distinguishing letter is used.

[0013] As used herein, various terms are used only for the purpose of describing particular embodiments and are not intended to be limiting. For example, the singular forms "a," "an," and "the" are intended to include the plural unless the context clearly indicates otherwise. Additionally, the terms "comprise," "comprises," and "comprising" are used interchangeably with "include," "includes," or "including." Additionally, the term "wherein" is used interchangeably with the term "where." As used herein, "exemplary" refers to examples, embodiments, and / or aspects and should not be considered as limiting or as indicating a preference or preferred embodiment. As used herein, order terms (such as "first," "second," "third," "Kth," "Nth," "Pth," etc.) used to modify elements such as structures, parts, operations, etc. do not themselves indicate a priority or order relative to another element, but merely distinguish the element from another element having the same name (but using order terms). As used herein, the term "set" refers to a group of one or more elements, and the term "plurality" refers to a plurality of elements.

[0014] As used herein, "generating," "calculating," "using," "selecting," "accessing," and "determining" are interchangeable unless the context indicates otherwise. For example, "generating," "calculating," or "determining" a parameter (or signal) can refer to actively generating, calculating, or determining a parameter (or signal), or can refer to using, selecting, or accessing a parameter (or signal) that has already been generated by another component, device, or the like.

[0015] As used herein, "coupled" can include "communicatively coupled," "electrically coupled," or "physically coupled," and further (or alternatively) can include any combination thereof. Two devices (or components) can be directly or indirectly coupled (e.g., communicatively coupled, electrically coupled, or physically coupled) via one or more other devices, components, wires, buses, networks (wired networks, wireless networks, or combinations thereof), and the like. Two devices (or components) that are electrically coupled can be included in the same device or in different devices and can be connected by electronics, one or more connectors, or inductive coupling, as illustrative, non-limiting examples. In some embodiments, two devices (or components) that are communicatively coupled, such as by electrical communication, can send and receive electrical signals (digital or analog signals) directly or indirectly via one or more wires, buses, networks, and the like. As used herein, "directly coupled" is used to describe two devices that are coupled (communicatively coupled, electrically coupled, or physically coupled) with no intervening parts.

[0016] 1 is a diagram illustrating an example system 100 that includes one or more aircraft 102 and an anomaly detection system 104. The aircraft 102 may be a rotorcraft 102A, a fixed-wing aircraft 102B, or other aircraft. The aircraft 102 includes aviation equipment 106, such as one or more components of a propulsion system, one or more components of a hydraulic system, one or more components of an electrical system, and one or more components of an environmental system.

[0017] The aircraft 102 further comprises one or more sensors 108 associated with the avionics 106. The sensors 108 are configured to generate sensor data 114. The sensor data 114 includes a plurality of data samples in a time series 116. Each data sample is indicative of a corresponding parameter value in the time series 116. For example, in FIG. 1, the time series 116 of data samples includes a first parameter value 118 ("PV 1 "), the second parameter value 120 (displayed as "PV 2 "), and the parameter value 122 of P (displayed as "PV p "). The first parameter value 118 corresponds to a value sensed by a particular one of the sensors 108 during a first sample period, the second parameter value 120 corresponds to a value sensed by a particular one of the sensors 108 during a second sample period, and the Pth parameter value 122 corresponds to a value sensed by a particular one of the sensors 108 during the Pth sample period (P is an integer equal to or greater than 2). The sample periods are sequential. For example, the Pth sample period follows a second sample period, which follows a first sample period. The time series of data samples 116 may further include one or more additional parameter values ​​(e.g., between the second parameter value 120 and the Pth parameter value 122), which are not shown for ease of illustration. The particular measurements or information represented by each of the parameter values ​​118-122 depend on the type of sensor 108 that generated the sensor data 114. For example, the parameter values ​​118 - 122 may indicate temperature, pressure, rotation rate, speed, position, humidity, vibration rate, or any other information related to the operation or status of the aircraft equipment 106 .

[0018] The sample periods of the time series 116 may be evenly spaced (e.g., periodically) or irregularly spaced (e.g., non-periodically). For example, one of the sensors 108 may collect data samples one every 10 milliseconds or other sampling rate to generate evenly spaced data samples. As another example, one of the sensors 108 may collect one data sample each time a state of the avionics device 106 changes (or changes by a particular amount) to generate irregularly spaced data samples of the time series 116. As yet another example, one of the sensors 108 may collect data samples at different sampling rates at different times, e.g., a first sampling rate during flight and a second sampling rate upon power-up on the ground.

[0019] In some embodiments, the sensor data 114 includes data samples from two or more sensors 108. In such embodiments, the sensor data 114 may include two or more time series 116 (e.g., one per sensor), and each of the two or more time series 116 may be evenly spaced or irregularly spaced. If the two or more time series 116 are both evenly spaced, the sampling rates associated with the two or more time series 116 may be equal to one another (e.g., the sampling rate for both sensors may be 15 Hz), or the sampling rates associated with the two or more time series 116 may be irregular (e.g., the sampling rate for one sensor may be 15 Hz and the sampling rate for the other sensor may be 20 KHz).

[0020] In some embodiments, in addition to the time series 116 of data samples indicative of the parameter values ​​118-122, the sensor data 114 further includes metadata. For example, the sensor data 114 may include time index information (e.g., a timestamp or sequence number) associated with the time series 116, a sensor identifier indicating which of the multiple sensors 108 generated the sensor data 114, and other information related to the sensor 108, the sensor data 114, the aircraft 102, the avionics, or the time series 116. The sensor data 114 may be represented by a digital signal or an analog signal. For example, the sensor data 114 may be transmitted as one or more digitally encoded packets transmitted via a digital communication protocol. As another example, the sensor data 114 may be encoded by one or more analog electrical signals, where a signal characteristic (e.g., magnitude, frequency, or phase) of the analog electrical signal indicates the parameter value.

[0021] The sensor data 114 is provided to the anomaly detection system 104, which is configured to analyze the sensor data 114 to predict or detect operational anomalies related to the aircraft equipment 106. In FIG. 1, the anomaly detection system 104 is shown as separate from the aircraft 102. For example, in FIG. 1, the anomaly detection system 104 and the aircraft 102 communicate over the network 124. However, in some embodiments, the anomaly detection system 104 is integrated with the aircraft 102. For example, the anomaly detection system 104 may be included in a health monitoring system (such as health monitoring system 816 shown in FIG. 8 ) of the aircraft 102.

[0022] 1, the anomaly detection system 104 includes, is contained within, or corresponds to a computing device. For example, the anomaly detection system 104 includes one or more processors 128, one or more memory devices 130, and an interface 132 (such as a data port, a wireless communication system, or other communication interface or input / output interface). The interface 132 is configured to receive the sensor data 114 and to transmit an output 168 generated by the anomaly detection system 104 to the on-board equipment 110 of the aircraft 102, to the ground equipment 126 (e.g., a maintenance computer system), or both. In other embodiments, the anomaly detection system 104 is implemented using application specific circuitry (e.g., an application specific integrated circuit), a programmed field programmable gate array, or a combination of the above.

[0023] In the example shown in FIG. 1, the memory device 130 stores instructions 134 executable by the processor 128 to analyze the sensor data 114 and generate the output 168. In the example shown in FIG. 1, the instructions 134 include various functional modules, including a delta value calculator 138, a delta value quantizer 140, a delta value counter 142, a normalization factor calculator 144, a counter normalizer 146, an anomaly detector 148, and an output generator 150. In the following description, the functional modules of the instructions 134 are described to illustrate one particular embodiment of the anomaly detection system 104. In other embodiments, the instructions 134 include more functional modules, fewer functional modules, or different functional modules. For example, in some embodiments, two or more of the functional modules are combined in a single module. By way of example, the delta value quantizer 140 and the delta value counter 142 can be combined in a histogram generator (not shown). In another example, one or more of the functional modules can be split into two or more separate modules. To illustrate, the anomaly detector 148 may be split into two separate anomaly detectors that use different detection thresholds, such as a first detection threshold for generating an alert (e.g., a maintenance alert) to be sent to the ground equipment 126 and a second detection threshold for generating an alert (e.g., an alert to the aircrew) to be sent to the on-board equipment 110. As yet another example, in some embodiments, one or more of the functions described below may be performed using dedicated hardware rather than software instructions. To illustrate, a dedicated circuit may be used to perform the operations of the delta value calculator 138 by delaying a sample and subtracting the sample from a subsequent sample.

[0024] In the example shown in FIG. 1 , the delta value calculator 138 is executable by the processor 128 to determine delta values ​​160 based on the sensor data 114. Each delta value 160 corresponds to the difference between parameter values ​​of two successive data samples in the time series 116. Illustratively, a first delta value ("DV 1 ") is the formula DV 1 =PV 2-PV 1 The delta value of P-1 (DV P-1 ") is a P-1 =PV P -PV P-1 In some embodiments, the delta value calculator 138 ignores, omits, or eliminates the sign of each delta value 160. For example, the first delta value (DV 1 ) is the formula DV 1 =|PV 2 -PV 1 In this example, each delta value 160 indicates the magnitude of the difference between adjacent parameter values ​​in the time series 116.

[0025] The delta value quantizer 140 is executable by the processor 128 to assign the delta values ​​160 to quantization bins (also referred to herein simply as "bins") to generate quantized delta values ​​162. Each bin represents a range of delta values, and the particular range of delta values, indicated by each bin and the number of bins, used for the sensor data 114 from each sensor 108 is indicated by bin data 154 in the configuration data 136 stored in the memory device 130. The bin data 154 is determined by a training process described with reference to FIGS. 3 and 5. In an example of assigning the delta values ​​160 to bins, the bin data 154 may specify evenly spaced bins, with each bin corresponding to a delta value range of 5 units. In this example, the particular delta value (e.g., the first delta value DV in the example above) may be assigned to a quantization bin 154. 1 ) is 4.5, then the particular delta value is assigned to the first bin (corresponding to a range of 0-5 units), and if the value of a particular delta value is 32.1, then the particular delta value is assigned to the seventh bin (corresponding to a range of 30-35 units).

[0026] In some embodiments, the bin data 154 implicitly or explicitly specifies a difference threshold, and only delta values ​​that meet the difference threshold are assigned to a quantization bin. To illustrate, the bin data 154 may indicate that the range for a first bin is 0.5 to 6. In this illustrative example, delta values ​​160 less than 0.5 are not assigned to a bin, thus resulting in a difference threshold of 0.5. Alternatively, the delta value calculator 138 may omit or zero out delta values ​​160 that do not meet the difference threshold.

[0027] The delta value counter 142 is executable by the processor 128 to determine a count of a number of delta values ​​160 assigned to each bin to generate a quantized delta value count 164. The quantized delta value count 164 for each bin indicates the number of times that the sensor data 114 changed by a particular amount between two consecutive samples during a particular power-up event.

[0028] The number of delta values ​​of a particular magnitude is expected to be larger for power-up events of long duration. In this context, a "power-up event" refers to an instance in which an engine or auxiliary power unit of the aircraft 102 is operating and powering the aviation equipment 106. A power-up event ends when an engine or auxiliary power unit of the aircraft 102 stops operating or stops powering the aviation equipment 106. The duration of a power-up event of the aircraft 102 may vary significantly, for example, from a few minutes during a short duration test to several hours during a long-distance flight. Typically, a power-up event of long duration is expected to have a larger quantized delta value count 164 for any particular bin than a power-up event of short duration. The anomaly detection system 104 normalizes the quantized delta value count 164 to a reference (e.g., normalization reference 152) to ensure that comparison to the anomaly detection threshold 156 is meaningful. Normalization adjusts the quantized delta value count 164 for each bin to an estimated value (eg, normalized count of delta values ​​166 ) that corresponds to a typical power-up duration, as indicated by the normalization criteria 152 of the configuration data 136 .

[0029] Normalization metric 152 indicates a benchmark duration of a typical power-up event or the number of data samples (e.g., a count of parameter values) that each sensor 108 transmits during a typical power-up event. For example, if a particular sensor 108 has a sampling rate of 1 Hz and a typical power-up event duration is 90 minutes, normalization metric 152 for a particular sensor 108 may indicate that a typical power-up event duration is 90 minutes, or that a typical power-up event duration is 5,400 samples (i.e., 1 sample / second) from a particular sensor 108. * 60 seconds / minute * 90 minutes). For a sensor 108 that produces evenly spaced data samples, the power-up event duration is equal to the normalization criterion 152 divided by the sampling rate of the sensor 108, and thus, for a sensor 108 that produces evenly spaced data samples, the normalization criterion 152 may represent either value. For a particular sensor 108 that produces irregularly spaced data samples, the duration of a typical power-up event is used as the normalization criterion 152. Another sensor 108 may have a different sampling rate or a different sampling pattern (e.g., periodic or aperiodic). Thus, each sensor 108 used for anomaly detection may be associated with a respective normalization criterion 152.

[0030] The duration of a typical power-up event is an arbitrary specified value, such as a value selected by a user associated with anomaly detection system 104 during or prior to training of anomaly detector 148. By way of example, the duration of a typical power-up event may be set based on an average duration of power-up events for a set of aircraft 102.

[0031] The normalization factor calculator 144 is executable by the processor 128 to determine a normalization factor based on the normalization criterion 152 and information describing a power-up event (e.g., power-up duration 158) associated with the sensor data 114. For example, the normalization factor may be determined by dividing the power-up duration 158 (or a count of sample periods during a power-up event) associated with the time series 116 by the duration of a typical power-up event (or a count of sample periods during a typical power-up event), as indicated by the normalization criterion 152.

[0032] The count normalizer 146 is executable by the processor 128 to normalize the counts 164 of the quantized delta values ​​based on a normalization factor to generate normalized counts 166 of the quantized delta values ​​(also referred to herein as normalized counts of delta values, or normalized counts). The normalized counts 166 of the quantized delta values ​​for a particular power-up event include one normalized count for each bin of each sensor 108 that provides sensor data 114 to the anomaly detection system 104. The normalized count for a particular bin represents an estimate (e.g., an extrapolation) of how many delta values ​​would be associated with the particular bin if the duration of the particular power-up event was equal to the duration of a typical power-up event.

[0033] The anomaly detector 148 is executable by the processor 128 to compare each of the normalized counts 166 of the delta values ​​to each of the anomaly detection thresholds 156. If the normalized counts 166 of the delta values ​​meet (e.g., are greater than or equal to) the respective anomaly detection thresholds 156, the anomaly detector 148 causes the output generator 150 to generate an output 168 indicating that an anomaly has been detected. The anomaly detection thresholds 156 are established during a training process, such as the training process described with reference to FIGS. 3 and 5.

[0034] The anomaly detection thresholds 156 include one or more thresholds for each bin of each sensor 108 that provides sensor data 114 to the anomaly detection system 104. For example, in a particular embodiment where only one sensor 108 provides sensor data 114 to the anomaly detection system 104, the anomaly detection thresholds 156 include one threshold for each bin, and the anomaly detection system 104 detects an anomaly if the normalized count of delta values ​​166 for any bin meets the corresponding anomaly detection threshold 156. As another example, in a particular embodiment where only one sensor 108 provides sensor data 114 to the anomaly detection system 104, the anomaly detection thresholds 156 include two thresholds for each bin (e.g., a first threshold and a second threshold). In this example, the anomaly detection system 104 detects a first type of anomaly if the normalized count of delta values ​​166 for any bin meets the corresponding first threshold, and detects a second type of anomaly if the normalized count of delta values ​​166 for any bin meets the corresponding second threshold. Illustratively, a first threshold may be used to trigger a notification to the ground equipment 126 and a second threshold may be used to trigger a notification to the on-board equipment 110 .

[0035] As yet another example, in certain embodiments, multiple sensors 108 provide respective sensor data 114 to the anomaly detection system 104, and the sensor data 114 from each sensor 108 is used to generate a respective normalized count 166 of delta values. In this example, the anomaly detection thresholds 156 include one or more anomaly detection thresholds 156 for each bin of the normalized count 166 of delta values ​​for each sensor 108. Thus, if 12 sensors 108 provide respective sensor data 114 to the anomaly detection system 104, then 12 normalized counts 166 of delta values, each having 10 bins, are generated, and the anomaly detection thresholds 156 specify at least 120 anomaly detection thresholds 156 (corresponding to 12 sensors x 10 bins per sensor). The number of bins used to quantize the delta values ​​may vary for each sensor 108. For example, ten bins may be used to quantize the delta values ​​160 associated with the sensor data 114 from a first sensor, and seven bins (or another number greater than or less than ten) may be used to quantize the delta values ​​160 associated with the sensor data 114 from a second sensor. Additionally, some sensors 108 may be associated with one anomaly detection threshold per bin, while other sensors 108 may be associated with two or more anomaly detection thresholds per bin.

[0036] In some embodiments, the anomaly detection thresholds 156 further include two separate thresholds for each bin of each sensor, with the particular thresholds used depending on the conditions present during the power-up event. For example, the anomaly detection system 104 may receive data identifying conditions present during the power-up event from one or more of the sensors 108, from a user (e.g., via the on-board equipment 110), or from another device (e.g., the ground equipment 126). In this example, the anomaly detection system 104 may select the particular anomaly detection thresholds 156 for each bin and each sensor 108 based on the data identifying the conditions present during the power-up event. To illustrate, the environmental control system of the aircraft 102 may operate differently depending on the ambient environmental conditions, such as the ambient temperature, ambient humidity, ambient pressure, etc. In this illustrative example, the anomaly detection thresholds 156 may include, for example, a first set of thresholds used when relatively warm weather conditions exist and a second set of thresholds used when relatively cold weather conditions exist. In this illustrative example, the anomaly detection system 104 can select whether to use the first set of thresholds or the second set of thresholds based on user input (e.g., via the on-board equipment 110), based on weather data (e.g., from the ground equipment 126), or based on sensor data 114 from one of the sensors 108 (e.g., an ambient temperature sensor).

[0037] In certain embodiments, some sensors 108 have many relatively small deviations during normal operating conditions (e.g., when there is no imminent failure). In such embodiments, some bins of the normalized count 166 of delta values ​​may not be associated with an anomaly detection threshold 156. For example, if normal operation of the avionics 106 monitored by a particular sensor 108 is associated with a sample-to-sample deviation of up to 5-10 units (e.g., a delta value of 5-10 units), then one or more bins associated with delta values ​​of 0-10 units are not associated with a corresponding anomaly detection threshold 156.

[0038] When an operational anomaly is detected by anomaly detector 148, output generator 150 generates output 168. Output 168 may include an anomaly notification message, a signal (e.g., voltage or current) provided to on-board display 112, etc. In some embodiments, output 168 may include details of the detected anomaly, such as an identification of the avionics 106 associated with the anomaly, an identification of the sensor 108 that generated the sensor data 114, a sample of the sensor data 114, a summary of the sensor data 114, a tag or other identifier associated with the anomaly detection threshold 156 that was met, etc.

[0039] The output 168 may be transmitted to the on-board equipment 110, to the ground equipment 126, or both. For example, the output 168 may be transmitted to the on-board equipment 110 such that the on-board indicators 112 notify crew members (e.g., flight crew or ground crew) of the operational anomaly. As another example, the ground equipment 126 may include or correspond to maintenance computer equipment or ground operations computer equipment. In this example, the output 168 may cause the ground equipment 126 to notify ground crew or maintenance crew of the operational anomaly or to automatically schedule one or more maintenance actions based on the operational anomaly.

[0040] By detecting operational anomalies before a fault condition occurs, the anomaly detection system 104 allows operators of the aircraft 102 to more effectively and efficiently plan maintenance for the aircraft 102. To detect anomalies, the anomaly detection system 104 uses a computationally efficient process similar to generating a histogram. For many sensor data columns, a large deviation over a short period of time is abnormal and may indicate an impending fault condition. By way of example, in an environmental control system, it is normal and expected to experience several relatively small deviations in compressor outlet pressure from time to time. However, a similar number of large deviations in compressor outlet pressure over the same period of time may indicate that the compressor (or pressure sensor) is beginning to experience conditions that typically lead to failure. By way of example, such wide deviations may indicate that the compressor is having difficulty starting to rotate the impeller, which may indicate the early stages of a bearing failure. Using the computationally efficient calculations and counting operations described above, the anomaly detection system 104 generates histogram-like data (e.g., normalized counts of delta values ​​166) that can be used to determine whether the sensor data 114 includes worrisome deviations, such as too many (e.g., at least a threshold number of) changes of a particular magnitude (e.g., in a particular bin). The use of histogram-like data also allows the anomaly detection system 104 to consider various ranges of deviations in the sensor data 114 and corresponding thresholds. For example, while a small number of "large" delta values ​​160 may indicate an anomaly, in some cases an anomaly may also (or alternatively) be indicated by a larger number of "medium" magnitude delta values ​​160. To illustrate, in some embodiments, 25 compressor outlet pressure delta values ​​in the range of 20-30 pounds per square inch (psi) may be just as worrisome as 4 compressor outlet pressure delta values ​​in the range of 40-50 psi over the same time period. The use of the histogram-like data described above allows the anomaly detection system 104 to effectively detect either condition as well as other conditions.

[0041] Figure 2 is a diagram illustrating details of operations 200 performed by anomaly detection system 104 of Figure 1 to detect anomalies based on sensor data 114. In certain embodiments, operations 200 described with reference to Figure 2 are performed by one or more processors 128 of Figure 1 executing functional modules of instructions 134 in accordance with configuration data 136. Operations 200 are performed during or after a power-up event of aircraft 102.

[0042] 2, the calculate delta values ​​operation 238 generates a set of delta values ​​160 based on the sensor data 114. For example, the calculate delta values ​​operation 238 determines the delta values ​​160 by calculating the difference, or the absolute value of the difference, between parameter values ​​indicated by each pair of successive data samples of the sensor data 114. Each delta value 160 thus indicates how much a measured parameter changes from one sample period of the sensor data 114 to the next.

[0043] In FIG. 2, the quantization operation 240 of the delta values ​​assigns each delta value 160 to a quantization bin 254 based on the magnitude of the respective delta value 160. For example, in FIG. 2, the first set of delta values ​​160 have magnitudes within the range associated with the first bin 254A and are assigned to the first bin 254A as the first set of quantized delta values ​​162A. Similarly, the Nth set of delta values ​​160 have magnitudes within the range associated with the Nth bin 254N and are assigned to the Nth bin 254N as the Nth set of quantized delta values ​​162N, where N is an integer equal to or greater than 2. The bin data 154 may indicate the number of bins 254 used, the range associated with each bin 254, or other information describing the bins 254. For example, the first bin data 154A may indicate the range of magnitudes associated with the first bin 254A, and the Nth bin data 154N may indicate the range of magnitudes associated with the Nth bin 254N.

[0044] 2, the delta value counting operation 242 counts the number of delta values ​​160 assigned to each bin 254. For example, the delta value counting operation 242 counts the number of delta values ​​160 in a first set of quantized delta values ​​162A to determine a first count 164A. Similarly, the delta value counting operation 242 counts the number of delta values ​​160 in an Nth set of quantized delta values ​​162N to determine an Nth count 164N.

[0045] 2, the count normalization operation 246 uses a normalization factor 210 to determine the normalized counts 164 of the quantized delta values. For example, the first count 164A associated with the first bin 254A is multiplied by the normalization factor 210 to determine the first normalized count 166A associated with the first bin 254A. Similarly, the Nth count 164N associated with the Nth bin 254N is multiplied by the normalization factor 210 to determine the Nth normalized count 166N associated with the Nth bin 254N. The normalization factor 210 is determined by the normalization factor operation 244 based on the power-up duration 158 and the normalization criterion 152. For example, in the embodiment shown in FIG. 2, the normalization factor 210 is equal to the power-up duration 158 divided by the normalization criterion 152.

[0046] In FIG. 2, the anomaly detection operation 248 compares one or more of the normalized counts 166 of the delta values ​​to one or more corresponding thresholds 156 to determine whether an abnormal condition is indicated. For example, a first normalized count 166A is compared to a first threshold 156A associated with a first bin 254A. Similarly, an Nth normalized count 166N is compared to an Nth threshold 156N associated with an Nth bin 254N. As described with reference to FIG. 1, in some embodiments, one or more of the bins 254 are not associated with a threshold 156. For example, if any number of delta values ​​160 within the range of delta values ​​associated with a particular bin are acceptable (e.g., not indicative of an imminent fault condition), then the particular bin is not associated with a threshold 156. Additionally, as described with reference to FIG. 1, in some embodiments, one or more of the bins 254 are associated with two or more thresholds 156. For example, a particular bin 254 may be associated with a first threshold (e.g., a lower threshold) and a second threshold (e.g., an upper threshold), and the output 168 generated may depend on the particular threshold that is met. As another example, a particular bin 254 may be associated with multiple thresholds, and the particular threshold 156 used by the anomaly detection operations 248 depends on conditions associated with a power-up event of the aircraft 102, such as whether the aircraft is flying during the power-up event, ambient environmental conditions associated with the power-up event, or particular functions or equipment operating during the power-up event.

[0047] 2, output generation operation 250 generates output 168 based on the comparison performed by anomaly detection operation 248. For example, the particular content of output 168, the location to which output 168 is sent, or both, may be selected by output generation operation 250 depending on which of one or more thresholds 156 have been met. In some embodiments, output 168 indicates that no anomaly has been detected in response to none of thresholds 156 being met.

[0048] Figure 3 is a diagram illustrating details of operations performed to train the anomaly detection system 104 of Figure 1. In certain embodiments, the operations 300 described with reference to Figure 3 are performed by one or more processors 128 of Figure 1 during training of the anomaly detection system 104 to generate at least a portion of the configuration data 136. In other embodiments, the operations 300 described with reference to Figure 3 are performed by another device separate from the anomaly detection system 104, such as another computing device. In such embodiments, the other device transmits at least a portion of the configuration data 136 to the anomaly detection system 104, or the anomaly detection system 104 retrieves at least a portion of the configuration data 136 from the other device.

[0049] The anomaly detection system 104 is trained based on sensor data 314 from multiple power-up events of one or more aircraft 102. In contrast to the sensor data 114 of Figures 1 and 2, which show a single power-up event of a particular aircraft 102, the sensor data 314 represents a history of multiple power-up events. For example, in Figure 3, a first set of sensor data 314A is generated during a power-up event of a first aircraft 102A, and a Kth set of sensor data 314K is generated during a power-up event of a Kth aircraft 102K, where K is an integer greater than or equal to 2. In some embodiments, one or more additional sets of sensor data 314 are generated during power-up events of one or more other aircraft 102. In some embodiments, multiple sets of sensor data 314 across multiple power-up events are generated by a single aircraft 102.

[0050] 3, the delta value calculation operation 338 generates a set of delta values ​​360 for each set of sensor data 314. For example, the delta value calculation operation 338 determines the first set of delta values ​​360A by calculating the difference, or the absolute value of the difference, between parameter values ​​indicated by each pair of consecutive data samples in the first set of sensor data 314A. Similarly, the delta value calculation operation 338 determines the Kth set of delta values ​​360K by calculating the difference, or the absolute value of the difference, between parameter values ​​indicated by each pair of consecutive data samples in the Kth set of sensor data 314K.

[0051] 3, the quantize delta values ​​operation 340 assigns the delta values ​​of each set of delta values ​​360 to a quantization bin 254 based on the magnitude of the respective delta value. For example, in FIG. 3, the first set of delta values ​​360A are assigned to bins 254 to generate a first set of quantized delta values ​​362A. Similarly, the Kth set of delta values ​​360K are assigned to bins 254 to generate a Kth set of quantized delta values ​​362K, where K is an integer equal to or greater than 2.

[0052] The number of bins 254 used, the range associated with each bin 254, and other information describing the bins 254 are determined based on the range of values ​​associated with the set of delta values ​​360, or based on the range of parameter values ​​indicated by each set of sensor data 314 or associated with the type of sensor that generated each set of sensor data 314. For example, the number of bins 254 used for a particular set of sensor data 314 can be user specified, can be a default value, can be selected based on heuristics, or can be optimized or determined based on a training process to give the anomaly detection system 104 specified detection characteristics. Typically, the bins 254 represent equally sized ranges. For example, if a first bin 254 represents a range of 5 units, then each of the other bins 254 also represents a range of 5 units.

[0053] When the number of bins 254 to be used is known, the range of each bin 254 can be determined by dividing the magnitude of the maximum delta value in the set of delta values ​​360 by the number of bins 254. In some embodiments, a safety margin can be added to the magnitude of the maximum delta value to ensure that the collective range of the bins 254 includes all possible delta values. Alternatively, a maximum delta value less than the particular magnitude can be used to split the bins 254 if the sensor data 314 indicates a fault condition when a delta value of a particular magnitude is detected.

[0054] If the number of bins 254 to be used is unknown or not specified in advance, the number of bins 254 may be selected arbitrarily or randomly during a first training iteration and may be updated (if necessary) during subsequent training iterations to improve the accuracy of the anomaly detection system 104. To illustrate, the anomaly detection system 104 may be trained using a set of training data (e.g., a set of sensor data 314 and fault indication data 308 associated with the aircraft 102 that generated the sensor data 314) and with a first number of bins 254. In this illustrative example, the fault indication data 308 indicates whether each particular aircraft 102 entered a fault condition after a power-up event associated with the sensor data 314 (e.g., within a particular number of subsequent power-up events). Thus, the fault indication data 308 acts as an indicator (for supervised training) of whether each set of sensor data 314 is associated with a future fault condition. Continuing with the illustrative example, after the anomaly detection system 104 has been trained using the first set of training data and the first number of bins 254, the anomaly detection system 104 may be tested using a set of test data to determine whether the anomaly detection system 104 can accurately predict which sensor data in the set of test data are indicative of a future fault condition. If the testing indicates that the anomaly detection system 104 meets the specified acceptance criteria (e.g., sufficient accuracy, an acceptably low rate of false positives, an acceptably low rate of false negatives), the anomaly detection system 104 may be released for use with the first number of bins 254. However, if the testing indicates that the anomaly detection system 104 does not meet the specified acceptance criteria, the anomaly detection system 104 may be retrained and retested using the second number of bins 254.

[0055] 3, the delta value counting operation 342 counts the number of delta values ​​assigned to each bin 254. For example, the delta value counting operation 342 counts the number of delta values ​​in the first set of quantized delta values ​​362A assigned to each of the bins 254 to determine a first bin count 364A, and counts the number of delta values ​​in the Kth set of quantized delta values ​​362K assigned to each of the bins 254 to determine a Kth bin count 364K.

[0056] 3, count normalization operation 346 determines normalized counts 366 based on bin counts 364 and normalization factors 310. Normalization factors 310 include one normalization factor for each set of sensor data 314. In other words, there are K normalization factors 310 corresponding to the K sets of sensor data 314. Each normalization factor 310 is equal to the power-up duration associated with the set of sensor data 314 divided by a normalization criterion. The normalization criterion is an arbitrary standardized power-up duration or a representative power-up duration, for example, the average power-up duration associated with all of the power-up events from which the set of sensor data 314 was generated.

[0057] 3, the threshold setting operation 312 generates one or more representative counts for each bin 254 based on the normalized counts 366. In the example illustrated in FIG. 3, the threshold setting operation 312 identifies a representative fault count 326 for each bin 254 based on an aggregate bin count (or an average aggregate bin count) associated with the bin 254 and the fault-indicating data 308. For example, the threshold setting operation 312 identifies a first representative fault count 326A for the first bin 254A. The first representative fault count 326A indicates a value of the normalized count 366 for the first bin 254A that may be indicative of an imminent fault condition. In this example, the representative fault counts 326A may be determined by a machine learning process, such as a support vector machine, that uses the set of normalized counts 366 and the fault-indicating data 308 as training data to identify, for each bin 254, a boundary between normalized counts 366 that do not indicate an imminent fault and normalized counts 366 that indicate an imminent fault. In other examples, a statistical or heuristic process may be used to identify, for each bin, a boundary between normalized counts 366 that do not indicate an imminent fault and normalized counts 366 that indicate an imminent fault. In some embodiments, the threshold setting operation 312 further identifies, for each bin 254, a representative non-fault count 324 using a similar process. The representative non-fault counts 324 indicate values ​​of the normalized counts 366 for the respective bin 254 that may not indicate an imminent fault condition.

[0058] In some embodiments, the representative fault count 326 associated with each bin 254 is set as the anomaly detection threshold 156 for the respective bin 254. For example, the representative fault count 326A associated with the first bin 254A is stored in the configuration data 136 as the first anomaly detection threshold 156A associated with the first bin data 154A, and the representative fault count 326N associated with the Nth bin 254N is stored in the configuration data 136 as the Nth anomaly detection threshold 156N associated with the Nth bin data 154N.

[0059] As discussed above, the anomaly detection system 104 can use more than one anomaly detection threshold for each bin 254. In certain embodiments, the threshold setting operation 312 uses the representative fault count 326 associated with each bin 254 to set the anomaly detection threshold 156 for the respective bin 254 higher and the representative non-fault count 324 associated with the bin 254 to set the anomaly detection threshold 156 for the bin 254 lower. Alternatively, the lower anomaly detection threshold for each bin 254 can be set based on an offset from the representative fault count 326 (e.g., by subtracting a particular value from the representative fault count 326, which may be arbitrary or may be based on statistics associated with the set of sensor data 314). In another alternative, the representative fault count 326 can be an average of the lowest normalized counts 366 that were associated with the fault conditions, and the high and low anomaly detection thresholds can be set based on a confidence interval or statistical variance of the normalized counts 366.

[0060] Once the anomaly detection thresholds 156 are established, the configuration data 136 may be used by the anomaly detection system 104 of FIG. 1 to detect an abnormal condition related to a power-up event in the aircraft 102 .

[0061] 4 is a flowchart of an example method 400 for detecting anomalies using the anomaly detection system 104 of FIG. 1. In certain embodiments, the method 400 is initiated, controlled, or performed by the processor 128 of FIG.

[0062] The method 400, at step 402, includes acquiring, with a computing device, sensor data captured by sensors of the aircraft during a power-up event. As a particular example, the anomaly detection system 104 of FIG. 1 receives sensor data 114, which is generated by one or more sensors 108 of the aircraft 102. The sensor data 114 may include a number of parameter values ​​(e.g., PV 1 118, PV 2 120, and P.V. P122), where each parameter value corresponds to a respective sample period during the power-up event.

[0063] The method 400 further includes determining, by a computing device, a set of delta values ​​at step 404. As a particular example, the delta value calculator 138 of FIGURE 1 calculates the delta values ​​160. Each delta value from the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods of the sensor data.

[0064] The method 400 further includes determining a set of quantized delta values ​​in the computing device by assigning the delta values ​​of the set of delta values ​​to quantization bins based on the magnitude of the delta values, at step 406. As a specific example, the delta value quantizer 140 of FIG. 1 assigns each delta value 160 to a quantization bin 254 based on the magnitude of the delta value 160 to generate a count 164 of quantized delta values.

[0065] The method 400 further includes determining, by the computing device, a normalized coefficient of the delta values ​​for each quantization bin, at step 408. As a particular example, the coefficient normalizer 146 of FIG. 1 determines the normalized coefficient of the delta values ​​for each of the quantization bins 254. Illustratively, the coefficient normalizer 146 multiplies the coefficient 164 of each quantized delta value for each quantization bin 254 by a normalization factor 210.

[0066] The method 400 further includes performing, by the computing device, a comparison of the normalized count of delta values ​​for the particular quantization bin to the anomaly detection threshold for the particular quantization bin at step 410. As a particular example, the anomaly detector 148 of FIG. 1 compares the normalized count of delta values ​​166 for the particular quantization bin 254 to the anomaly detection threshold 156 for the particular quantization bin 254. In some embodiments, multiple anomaly detection thresholds 156 are associated with some of the quantization bins 254. In such embodiments, the normalized count of delta values ​​166 is compared to each of the anomaly detection thresholds 156. Alternatively, in such embodiments, one of the anomaly detection thresholds 156 for the particular quantization bin 254 is selected (e.g., based on a condition regarding a power-up event) and the normalized count of delta values ​​166 for the particular quantization bin 254 is compared to the selected anomaly detection threshold 156.

[0067] Method 400 further includes generating, by the computing device and based on the comparison, an output indicative of whether the sensor data indicates an operational anomaly of the aviation equipment associated with the sensor, at step 412. As a particular example, output generator 150 of FIG. 1 generates output 168 that indicates whether an operational anomaly has been detected in the sensor data 114.

[0068] Method 400 is a computationally efficient process for detecting operational anomalies before a fault condition occurs, thereby enabling aircraft operators and maintainers to more effectively and efficiently plan maintenance.

[0069] Figure 5 is a flow chart of an example method 500 for determining an anomaly detection threshold for the anomaly detection system 104 of Figure 1. In certain embodiments, the method 500 is initiated, controlled, or performed by the processor 128 of Figure 1. In other embodiments, the method 500 is performed by a computing device separate from the anomaly detection system 104 (such as the ground device 126), which subsequently provides the anomaly detection threshold (and possibly other configuration data) to the anomaly detection system 104.

[0070] Method 500 includes, at step 502, acquiring with a computing device a plurality of sets of sensor data captured by on-board sensors of one or more aircraft during a plurality of power-up events, each set of sensor data corresponding to a respective power-up event of a respective aircraft and including a plurality of parameter values, each parameter value corresponding to a respective sample period during a respective power-up event.

[0071] The method 500 further includes determining, by the computing device, a set of delta values ​​for each set of sensor data, at step 504. Each delta value from the set of delta values ​​indicates a difference between a first parameter value and a second parameter value from a pair of parameter values ​​corresponding to successive sample periods from the set of sensor data. For example, the computing device may perform one or more delta value calculation operations 338, such as those described with reference to FIG.

[0072] The method 500 further includes determining, at a computing device, a set of quantized delta values ​​by assigning the delta values ​​of the set of delta values ​​to quantization bins based on the magnitudes of the delta values, at step 506. For example, the computing device may perform the quantization operation 340 of one or more delta values, as described with reference to FIG.

[0073] The method 500 further includes determining, by a computing device, normalized coefficients of the delta values ​​for each quantization bin, at step 508. For example, the computing device may perform one or more of the delta value counting operation 342 and the coefficient normalization operation 346, as described with reference to FIG.

[0074] The method 500 further includes setting an anomaly detection threshold for each quantization bin based on the normalized count of the delta values ​​and the fault indication data for the one or more aircraft, at step 510. For example, the computing device may perform one or more threshold setting operations 312, such as those described with reference to FIG.

[0075] FIG. 6 is a block diagram of a computing environment 600 including a computing device 610 configured to support aspects of the computer-implemented methods and computer-executable program instructions (or code) according to the present disclosure. As a particular example, the computing device 610 may be included within or correspond to the anomaly detection system 104 of FIG. 1. As another particular example, the computing device 610 may be included within or correspond to the ground device 126 or the airborne device 110 of FIG. 1. The computing device 610, or portions thereof, are configured to execute instructions to initiate, perform, or control one or more operations described with reference to FIGS. 1-5. For example, in some embodiments, the computing device 610 includes components configured to execute operations of the method 400 of FIG. 4, the method 500 of FIG. 5, or both. By way of example, in FIG. 6, the computing device 610 includes instructions 134 of FIG. 1, which may be executed to perform anomaly detection based on sensor data. In another illustrative example, in FIG. 6, computing device 610 comprises an anomaly detection trainer 636, which corresponds to instructions executable to perform operations 300, method 500, or both.

[0076] The computing device 610 includes one or more processors 620. The processors 620 are configured to communicate with a system memory 630, one or more storage devices 640, one or more input / output interfaces 650, one or more communication interfaces 660, or any combination thereof. The system memory 630 includes volatile memory devices (e.g., random access memory (RAM) devices), non-volatile memory devices (e.g., read only memory (ROM) devices, programmable read only memory, and flash memory), or both. The system memory 630 includes an operating system 632, which may include a basic input / output system for powering up the computing device 610, as well as a complete operating system that allows the computing device 610 to interact with a user, other programs, and other devices. The system memory 630 further stores applications, such as instructions 134, anomaly detection trainer 636, or both.

[0077] The one or more storage devices 640 include non-volatile storage devices, such as magnetic disks, optical disks, or flash memory devices. In certain examples, the storage devices 640 include both removable and non-removable memory devices. The storage devices 640 are configured to store an operating system, an image of the operating system, applications (e.g., one or more of the applications 634), and program data (e.g., the configuration data 136 of FIGS. 1-3). In certain aspects, the system memory 630, the storage devices 640, or both include tangible computer-readable media. In certain aspects, one or more of the storage devices 640 may be external to the computing device 610.

[0078] The one or more input / output interfaces 650 enable the computing device 610 to communicate with one or more input / output devices 670 to facilitate interaction with a user. For example, the one or more input / output interfaces 650 may include a display interface, an input interface, or both. The processor 620 is configured to communicate with a device or controller 680 via one or more communication interfaces 660. For example, the one or more communication interfaces 660 may include a network interface. The device or controller 680 may include, for example, the aircraft 102, devices of the network 124, sensors 108, on-board equipment 110, ground equipment 126, or any combination thereof.

[0079] In some embodiments, a non-transitory computer-readable medium stores instructions that, when executed by one or more processors, cause the one or more processors to initiate, execute, or control operations to perform some or all of the functions described above. For example, the instructions may be executable to perform one or more of the operations or methods of FIGS. 1-5. By way of example, the instructions of the application 634, when executed by the processor 620, cause the processor 620 to initiate, execute, or control operations to train the anomaly detection system 104, or to detect anomalies based on the sensor data 114, or both. In some embodiments, some or all of the operations or methods of one or more of FIGS. 1-5 are performed by one or more processors (e.g., one or more central processing units (CPUs), one or more graphics processing units (GPUs), one or more digital signal processors (DSPs)) executing instructions, or by dedicated hardware circuitry, or any combination thereof.

[0080] Figure 7 is a flow chart of an example lifecycle 700 of an aircraft equipped with the anomaly detection system of Figure 1. During preparation for production, the example lifecycle 700 includes, at step 702, the specification and design of an aircraft, such as aircraft 102. During the specification and design of the aircraft, the lifecycle 700 may include the specification and design of the anomaly detection system 104. At step 704, the lifecycle 700 further includes material procurement, which may include procuring materials for the anomaly detection system 104.

[0081] During production, the lifecycle 700 includes manufacturing of components and subassemblies at step 706 and system integration with the aircraft at step 708. For example, the lifecycle 700 may include manufacturing of components and subassemblies of the anomaly detection system 104 and system integration of the anomaly detection system 104. The lifecycle 700 includes certifying and delivering the aircraft at step 710 and placing the aircraft in service at step 712. The certification and delivery may include certifying the anomaly detection system 104 to place the anomaly detection system 104 in service. While in service with the customer, the aircraft may be scheduled for routine maintenance and service inspections (which may include refurbishment, reconfiguration, refurbishment, etc.). At step 714, the lifecycle 700 includes performing maintenance and service inspections on the aircraft, which may include performing maintenance and service inspections on the anomaly detection system 104. Alternatively, maintenance and service inspections may be performed on the aircraft based on operational anomalies detected by the anomaly detection system 104.

[0082] Each of the processes in lifecycle 700 may be performed or conducted by a system integrator, a third party, and / or an operator (e.g., a contractor). For purposes of this description, a system integrator may include, without limitation, any number of aircraft manufacturers and major system subcontractors, a third party may include, without limitation, any number of vendors, subcontractors, and suppliers, and an operator may be an airline, a leasing company, a defense company, a maintenance organization, etc.

[0083] Figure 8 is a block diagram of an aircraft 102. In the example shown in Figure 8, the anomaly detection system of Figure 1 is a component of the aircraft 102. In Figure 8, the aircraft 102 comprises an airframe 802 having a number of systems, such as aviation equipment 106 and an interior 804. The example aviation equipment 106 includes one or more components of a propulsion system 808, an electrical system 812, an environmental system 814, a hydraulic system 810, and a health monitoring system 816. In the example shown in Figure 8, the sensor 108, the anomaly detection system 104, or both, are part of the health monitoring system 816. Any number of other systems and other equipment may be included.

[0084] The diagrams of the examples described herein are intended to provide a general understanding of the structure of the various embodiments. The diagrams are not intended to be a complete description of all components and aspects of the apparatus and systems that use the structures or methods described herein. Many other embodiments will be apparent to those skilled in the art upon review of the present disclosure. Other embodiments may be utilized and obtained from the present disclosure, such that structural and logical substitutions and changes may be made without departing from the scope of the present disclosure. For example, the method operations may be performed in a different order than shown in the figures, or one or more method operations may be omitted. The present disclosure and the drawings are therefore to be regarded as illustrative, and not restrictive.

[0085] Moreover, while specific examples have been shown and described herein, it should be understood that any subsequent configurations designed to achieve the same or similar results may be substituted for the specific embodiments shown. The present disclosure is intended to cover any and all subsequent adaptations or variations of the various embodiments. Combinations of the above-described embodiments, as well as other embodiments not specifically described herein, will be apparent to those of skill in the art upon reading this description.

[0086] The Abstract of the Disclosure is appended with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. Also, in the foregoing Detailed Description of the Invention, various features may be grouped together or described in a single embodiment for the purpose of streamlining the disclosure. The foregoing examples are illustrative and do not limit the disclosure. It should also be understood that many modifications and variations are possible in accordance with the principles of the disclosure. As reflected in the following claims, the claimed subject matter may relate to less than all aspects of any disclosed example. Accordingly, the scope of the disclosure is defined by the following claims and their equivalents. [Explanation of symbols]

[0087] 100 Systems 102 Aircraft 104 Anomaly Detection System 106 Aviation equipment 108 Sensors 110 Onboard equipment 112 Onboard display 114 Sensor Data 116 Time Series 118 First Parameter Value 120 Second parameter value 122 Parameter value of P 124 Network 126 Ground Equipment 128 processors 130 Memory Device 132 Interface 134 Command 136 Configuration Data 138 Delta Value Calculator 140 Delta value quantizer 142 Delta value counter 144 Normalization Factor Calculator 146 Count Normalizer 148 Anomaly Detector 150 output generator 152 Normalization Criteria 154 Bin Data 156 Anomaly Detection Threshold 158 Power Up Duration 160 Delta value 162 Quantized Delta Values 164 Quantized delta value counting, normalized counting 166 Normalized coefficient of delta value

Claims

1. acquiring (402) with a computing device (104 or 610) sensor data (114) captured by a sensor (108) of the aircraft (102) during a power-up event, the sensor data including a plurality of parameter values ​​(118, 120, 122), each parameter value corresponding to a respective sample period during the power-up event; determining (404), by the computing device, a set of delta values ​​(160), each delta value from the set of delta values ​​indicating a difference between a first parameter value and a second parameter value of a pair of parameter values ​​corresponding to successive sample periods of the sensor data; determining (406), by the computing device, a set of quantized delta values ​​(162) by assigning delta values ​​from the set of delta values ​​to quantization bins (254) based on the magnitudes of the delta values; determining (408), by said computing device, for each of said quantization bins, a normalized coefficient (166) of delta values; performing (410) a comparison, by the computing device, of the normalized coefficient of the delta value for a particular quantization bin with an anomaly detection threshold (156) for the particular quantization bin; generating (412), by the computing device and based on the comparison, an output (168) indicative of whether the sensor data is indicative of an operational anomaly of an aviation instrument (106) associated with the sensor; Including, The power-up event refers to an instance in which an engine or auxiliary power unit of the aircraft (102) is operating to provide power to the aviation equipment (106), according to the method (400).

2. The method of claim 1 , further comprising the step of comparing, by the computing device, each delta value to a difference threshold, wherein only delta values ​​that satisfy the difference threshold are assigned to a quantization bin.

3. determining a power-up duration (158) of the power-up event indicated by the sensor data; determining a normalization factor (210) based on the power-up duration of the power-up event and based on a normalization criterion (152); determining a count (164) of delta values ​​assigned to each of the quantization bins; and adjusting the coefficients of the delta values ​​assigned to each of the quantization bins based on the normalization factor to determine normalized coefficients of the delta values. The method according to claim 1 or 2.

4. The method of claim 1 , further comprising selecting the anomaly detection threshold for each of the quantization bins based on an operating condition associated with the power-up event.

5. The method of claim 4 , wherein the aircraft devices correspond to components of an environmental system (814) and the operating conditions are related to ambient environmental conditions.

6. 6. The method of claim 1, wherein the anomaly detection threshold for each quantization bin comprises a first threshold and a second threshold, and the output comprises a first indication in response to the comparison indicating that a normalized count of delta values ​​associated with the particular quantization bin is greater than or equal to the first threshold associated with the particular quantization bin and less than the second threshold associated with the particular quantization bin.

7. 7. The method of claim 6, wherein the output includes a second indication responsive to the comparison indicating that the normalized count of the delta values ​​associated with the particular quantization bin is greater than or equal to the second threshold value associated with the particular quantization bin.

8. A non-transitory computer-readable storage medium (130) storing instructions (134) that, when executed by one or more processors (128), cause the one or more processors to: acquiring (402) sensor data (114) captured by a sensor (108) of the aircraft (102) during a power-up event, the sensor data including a plurality of parameter values ​​(118, 120, 122), each parameter value corresponding to a respective sample period during the power-up event; determining (404) a set of delta values ​​(160), each delta value from the set of delta values ​​indicating a difference between a first parameter value and a second parameter value of a pair of parameter values ​​corresponding to successive sample periods of the sensor data; determining (406) a set of quantized delta values ​​(162) by assigning delta values ​​from the set of delta values ​​to quantization bins (254) based on the magnitudes of the delta values; determining (408) a normalized coefficient (166) of delta values ​​for each of the quantization bins; performing (410) a comparison of the normalized coefficient of the delta value for a particular quantization bin with an anomaly detection threshold (156) for the particular quantization bin; generating (412) an output (168) indicative of whether the sensor data is indicative of an operational anomaly of an aviation instrument (106) associated with the sensor based on the comparison; Performing an action including The power-up event refers to an instance in which an engine or auxiliary power unit of the aircraft (102) is operating to provide power to the aviation equipment (106).

9. 10. The non-transitory computer-readable storage medium of claim 8, wherein the operations further comprise comparing each delta value to a difference threshold, and wherein only delta values ​​that satisfy the difference threshold are assigned to a quantization bin.

10. The operation, determining a power-up duration (158) of a power-up event indicated by said sensor data; determining a normalization factor (210) based on the power-up duration of the power-up event and based on a normalization criterion (152); determining a count (164) of delta values ​​assigned to each of the quantization bins; and adjusting the coefficients of the delta values ​​assigned to each of the quantization bins based on the normalization factor to determine normalized coefficients of the delta values.

10. A non-transitory computer readable storage medium according to claim 8 or 9.

11. 11. The non-transitory computer-readable storage medium of claim 8, wherein the operations further comprise selecting the anomaly detection threshold for each quantization bin based on an operating condition associated with the power-up event.

12. The step of performing the comparison comprises:

12. The non-transitory computer-readable storage medium of claim 8, further comprising: comparing a normalized count of the delta values ​​to a first threshold associated with the particular quantization bin and to a second threshold associated with the particular quantization bin, wherein the output includes a first indication in response to the comparison indicating that the normalized count of the delta values ​​associated with the particular quantization bin is greater than or equal to the first threshold and less than the second threshold.

13. 13. The non-transitory computer-readable storage medium of claim 12, wherein the output further comprises a second indication responsive to the comparison indicating that the normalized count of the delta values ​​is greater than or equal to the second threshold.

14. An aircraft (102), a sensor (108) for capturing sensor data (114) during a power-up event, the sensor data including a plurality of parameter values ​​(118, 120, 122), each parameter value corresponding to a respective sample period during the power-up event; one or more processors (128); a memory device (130) storing the sensor data and instructions (134) executable by the one or more processors to detect an operational anomaly of an aviation instrument (106) associated with the sensor, the instructions, when executed by the one or more processors, causing the one or more processors to: determining (404) a set of delta values ​​(160), each delta value from the set of delta values ​​indicating a difference between a first parameter value (118) and a second parameter value (120) of a pair of parameter values ​​corresponding to successive sample periods of the sensor data; determining (406) a set of quantized delta values ​​(162) by assigning delta values ​​from the set of delta values ​​to quantization bins (254) based on the magnitudes of the delta values; determining (408) a normalized coefficient (166) of delta values ​​for each of the quantization bins; performing (410) a comparison of the normalized coefficient of the delta value for a particular quantization bin with an anomaly detection threshold (156) for the particular quantization bin; generating (412) an output (168) indicative of whether the sensor data indicates an abnormality in the operation of the aviation equipment based on the comparison; A memory device (130) for performing operations including: Equipped with The power-up event refers to an instance in which an engine or auxiliary power unit of the aircraft (102) is operating to provide power to the aircraft equipment (106).

15. 15. The aircraft of claim 14, further comprising an on-board display (112), wherein the action further comprises transmitting the output to the on-board display.

16. 16. The aircraft of claim 14 or 15, wherein the operation further comprises transmitting the output to a ground device (126).

17. A method (500) for training an anomaly detection system (104), the method comprising: acquiring (502), with a computing device (610), a plurality of sets of sensor data (314) captured by on-board sensors (108) of one or more aircraft (102) during a plurality of power-up events, each set of sensor data corresponding to a respective power-up event of a respective aircraft and including a plurality of parameter values ​​(118, 120, 122), each parameter value corresponding to a respective sample period during the respective power-up event; determining (504), by the computing device, a set of delta values ​​(360) from each set of sensor data, each delta value in the set of delta values ​​indicating a difference between a first parameter value and a second parameter value of a pair of parameter values ​​corresponding to successive sample periods of the set of sensor data; determining (506), by the computing device, a set of quantized delta values ​​(362) by assigning the delta values ​​of the set of delta values ​​to quantization bins (254) based on the magnitudes of the delta values; determining (508), by the computing device, a normalized coefficient (366) of delta values ​​for each of the quantization bins; setting (510) an anomaly detection threshold (156) for each quantization bin based on the normalized coefficients of the delta values ​​and fault indication data (308) associated with the one or more aircraft; Including, The method, wherein the power-up event refers to an instance in which an engine or auxiliary power unit of the aircraft (102) is operating to provide power to an aviation device (106).

18. 20. The method of claim 17, wherein the multiple sets of sensor data include a first set of sensor data associated with a fault condition, and the anomaly detection threshold associated with a particular quantization bin is further set based on a normalized coefficient of a first delta value associated with the first set of sensor data.

19. 19. The method of claim 17 or 18, further comprising the step of comparing, by the computing device, each delta value with a difference threshold, wherein only delta values ​​that satisfy the difference threshold are assigned to a quantization bin.

20. determining, for each set of sample data, a power-up duration indicated by said respective set of sensor data; determining a normalization factor (210 or 310) for each set of sample data based on the power-up duration (158) indicated by the respective set of sensor data and based on a normalization criterion (152); determining a count of delta values ​​(360) assigned to each quantization bin of said respective set of sample data; adjusting the coefficients of the delta values ​​assigned to each quantization bin of the respective sets of sample data based on the normalization factor to determine a normalized coefficient of the delta values ​​for each quantization bin for the respective sets of sample data; 20. The method of any one of claims 17 to 19, comprising:

Images