System and method for improving maintenance planning in the presence of sensor data

Abstract

The present disclosure relates to systems and methods for improving maintenance plans in the presence of sensor data. Systems and methods for generating sensor-driven optimized and / or improved maintenance programs are provided. A first maintenance plan for a device type is received; the first maintenance plan defines maintenance tasks for the device type and defines intervals at which the maintenance tasks are to be performed. Historical sensor data, historical scheduled maintenance data, and historical unscheduled maintenance data are received and correlated to determine predicted values ​​for the sensor data. One or more maintenance improvement goals for operators of devices of the device type are received, and a sensor-driven optimized and / or improved maintenance program is generated based at least on the predicted values ​​of the sensor data and the maintenance improvement goals. The sensor-driven optimized and / or improved maintenance program includes adjusted intervals within the defined intervals.

Description

Technical Field

[0001] The present disclosure relates to systems and methods for improving maintenance planning in the presence of sensor data. Background Art

[0002] Fleet operators follow maintenance plans in order to provide efficient operations and fleet utilization. Examples of such vehicles include, but are not limited to, aircraft, air transport vehicles, automobiles, unmanned aerial vehicles (UAVs), marine vehicles (ships, submarines, etc.), etc. In the case of aircraft, maintenance planning data (MPD) is typically provided by the manufacturer to the operator of the aircraft and typically includes information related to the recommended scheduled maintenance tasks, their intervals, required visits, and other relevant information. Maintenance tasks are typically grouped and performed in packages identified as "check packages."

[0003] In some operational scenarios, operators use the MPD as a starting point and adjust their maintenance schedule based on financial considerations (e.g., the cost of removing a vehicle from service and performing a specific inspection package). However, maintenance schedule adjustments that are not informed by actual operating conditions and available historical data can result in maintenance being performed more frequently than necessary (thereby increasing costs) or less frequently than optimally, risking the operation being below a target acceptable maintenance schedule that leverages historical maintenance data and sensor data. Summary of the invention

[0004] The disclosed examples are described in detail below with reference to the accompanying drawings listed below. The following overview is provided to illustrate some examples disclosed herein. However, it is not intended to limit all examples to any specific configuration or sequence of operations.

[0005] Some aspects and implementations disclosed herein relate to methods for generating a sensor-driven optimized (or improved) maintenance program (SDOMP). The example method includes: receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed; receiving historical sensor data (including sensor alarm data), historical scheduled maintenance data, historical unscheduled maintenance data, and a document linking sensor alarms with maintenance tasks; associating the historical sensor data with the historical scheduled maintenance data and the historical unscheduled maintenance data to determine a predicted value of the sensor data; receiving one or more maintenance improvement goals of an operator of a device of the device type; and generating an SDOMP based at least on the predicted value of the sensor data and the maintenance improvement goal, the generated SDOMP including an adjusted interval in the defined interval.

[0006] Also provided herein are systems and methods for generating an SDOMP. A first maintenance plan for a device type is received; the first maintenance plan defines maintenance tasks for the device type and defines intervals at which the maintenance tasks are to be performed. Historical sensor data, historical scheduled maintenance data, and historical unscheduled maintenance data are received and correlated to determine predicted values ​​for the sensor data. One or more maintenance improvement goals for operators of devices of the device type are received, and an SDOMP is generated based on at least the predicted values ​​for the sensor data and the maintenance improvement goals. The SDOMP includes adjusted intervals within the defined intervals. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The disclosed examples are described in detail below with reference to the following figures:

[0008] Figure 1 It is a block diagram of the device production and service method.

[0009] Figure 2 is a block diagram of an apparatus 200 for use with various aspects of the present disclosure.

[0010] Figure 3 is a schematic perspective view of a particular flying device 201 .

[0011] Figure 4 is a block diagram of an implementation of a storage area 410 of a computing device 400 .

[0012] Figure 5 Overlay probability density curves and thresholds driving detection probability and false alarm probability values ​​are shown.

[0013] Figure 6 The cumulative distribution function (CDF) for the probability of detecting a condition requiring repair or other maintenance action is shown.

[0014] Figure 7 Shows Figure 6 The CDF curve of the changed threshold.

[0015] Figure 8 Illustration of how to determine Figure 7 The threshold of change.

[0016] Fig. 9 is a block diagram of an implementation of a process flow for implementing aspects of the present disclosure.

[0017] Fig.10 It is to realize the alarm display.

[0018] Fig.11 is a block diagram of an implementation of a sensor mapping operation.

[0019] Fig.12 is a graphical representation of a timeline of relevant events associated with generating SDOMP 430 .

[0020] Fig.13 The relationship between non-routine discovery and maintenance messages is shown.

[0021] FIG. 14A to FIG. 14D The net benefits of various scenarios utilizing sensor data are shown.

[0022] Fig.15 is a flowchart illustrating implementation of an operation of determining whether to add a maintenance task.

[0023] Fig.16 is a flow chart illustrating an implementation of the operation of generating SDOMP 430.

[0024] Fig.17 is a block diagram illustrating a process for generating SDOMP 430 .

[0025] Fig.18 is a block diagram illustrating an implementation of a system suitable for implementing various aspects of the present disclosure.

[0026] Corresponding reference characters indicate corresponding parts throughout the several views. DETAILED DESCRIPTION

[0027] Various embodiments will be described in detail with reference to the accompanying drawings. Whenever possible, the same reference numerals are used throughout the drawings to represent the same or similar components. References to specific implementations and implementations throughout this disclosure are provided for illustrative purposes only, but are not intended to limit all implementations unless otherwise indicated.

[0028] When read in conjunction with the accompanying drawings, the above overview and the following detailed description of certain embodiments will be better understood. As used herein, an element or step referenced in the singular and preceded by the word "one" or "an" should be understood to not necessarily exclude multiple elements or steps. In addition, reference to "an embodiment" is not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the stated features. In addition, unless explicitly stated to the contrary, an embodiment that "comprises" or "has" one or more elements with a particular characteristic can include additional elements that do not have that characteristic.

[0029] Sensors on the equipment are used for the purpose of developing sensor-based predictive tools to predict failures. This has the potential to replace preventive maintenance with predictive maintenance. However, sensors are often deployed by component suppliers or design engineers who usually do not have full knowledge of the inspection procedure. Therefore, functional mapping between sensors and maintenance tasks enables inspection intervals to be improved through virtual (sensor-enabled) inspections. Unfortunately, due to the probabilistic nature of some sensor data, some operators may ignore alarms and assume them to be false alarms. This hinders the realization of the full potential benefits of sensors. Typically, sensor alarms are a subset of sensor data because some alarms are binary conditions of sensor measurements exceeding a threshold.

[0030] Probabilistic sensor data can be converted into objective, rigorous and actionable information in the form of modified inspection intervals that do not rely solely on operator interpretation of alarms. Thus, when sensors are at least partially predictors of failure or degradation, the benefits of sensors are realized through reduced maintenance costs, improved efficiency and increased plant utilization.

[0031] Aspects of the present disclosure provide a computer-implemented method, apparatus, and computer program product for generating a sensor-driven optimized maintenance program (SDOMP). A first maintenance plan for an apparatus type is received; the first maintenance plan defines maintenance tasks for the apparatus type and defines intervals at which the maintenance tasks are to be performed. Historical sensor data, historical scheduled maintenance data, and historical unscheduled maintenance data are received and correlated to determine predicted values ​​for the sensor data. One or more maintenance improvement goals of an operator of an apparatus of the apparatus type (e.g., an owner of the apparatus or a fleet of the apparatus, an administrator responsible for the apparatus or a fleet of the apparatus, or a user operating the apparatus) are received, and an SDOMP is generated based at least on the predicted values ​​of the sensor data and the maintenance improvement goals. The SDOMP includes adjusted intervals within the defined intervals.

[0032] For a particular device (e.g., a type or model of a particular device or device fleet), the historical maintenance data includes the costs incurred for each maintenance event under an approved maintenance plan (AMP). The historical maintenance data includes historical scheduled maintenance data and historical unscheduled maintenance data. A maintenance event is any event associated with the maintenance, repair, or replacement of a component of a device. In one embodiment, maintenance events include, for example, functional component failure, system failure, loss of function, reduced function, service interruption, corrosion, wear, slow response time, reduced efficiency, reduced fuel efficiency, loss of tire pressure, or any other event that requires maintenance, repair, or replacement of a component or a sub-part of a component. In addition, each maintenance event includes one or more maintenance packages, and each maintenance package includes one or more maintenance tasks. As used herein, a maintenance task is a task associated with inspecting, maintaining, repairing, and / or replacing a component or sub-component. Using historical maintenance data, maintenance events including one or more maintenance packages are analyzed to determine the costs associated with each maintenance task.

[0033] Using sensor data and historical maintenance data, costs, and one or more maintenance improvement goals (e.g., provided by the operator of the device or fleet of devices), available resources (e.g., worker and facility availability at any given time), and a defined time length for the total operational life cycle of the device (e.g., a time length defined by the operator who owns the device or fleet of devices), each of a plurality of maintenance tasks, aspects of the present disclosure minimize the frequency of planned maintenance events, and improve device scheduling by optimizing or improving scheduling of preventive planned maintenance tasks while minimizing the occurrence of maintenance tasks in unplanned services. For example, by increasing the interval between maintenance tasks (e.g., the number of days between performing maintenance tasks), the operator who owns the device or fleet of devices (e.g., based on a lease agreement) performs a specific maintenance task on the device or fleet of devices The number of times within the total operational life cycle is reduced. By postponing maintenance of a specific maintenance task, the cost associated with performing the specific maintenance is also postponed. For a leased device or fleet of devices, the increased interval can cause the specific maintenance to be postponed beyond the lease return date of the device. In this way, assuming that the device is returned before the specific maintenance is performed, the equipment company will not bear the cost of the specific maintenance.

[0034] Scheduled maintenance activities are those that are planned in advance, while unscheduled maintenance activities are those that occur due to a malfunction or non-routine finding. Routine finding is a condition that is detected or determined during an inspection (whether scheduled or unscheduled) as expected, such as the tread depth on a tire after a certain amount of use. Non-routine finding is an unexpected condition that is detected or determined during an inspection (whether scheduled or unscheduled), such as corrosion or degradation of certain components in a state that is beyond the expected state and / or loss of function.

[0035] More specifically, referring to the accompanying drawings, Figure 1 The device manufacturing and service method 100 shown in FIG. Figure 2 Embodiments of the present disclosure are described in the context of the apparatus 200 provided in Figure 1 , shows a diagram of a method for manufacturing and servicing a device (eg, device 200) according to an embodiment. In one embodiment, during pre-production, the device manufacturing and servicing method 100 includes Figure 2 The specification and design 102 of the device 200 and the material procurement 104 are carried out during production. Figure 2 Components and sub-components of the device 200 are manufactured 106 and system integrated 108. Thereafter, Figure 2 The device 200 in the embodiment undergoes authentication and delivery 110 to be placed in service 112. While serving the customer, Figure 2 The device 200 in is scheduled for routine maintenance and service 114, which in one embodiment includes modification, reconfiguration, refurbishment, and other maintenance or service as described herein.

[0036] In one embodiment, each process of the device manufacturing and service method 100 is performed or completed by a system integrator, a third party, and / or an operator. In these implementations, the operator is a customer. For the purpose of description, the system integrator includes any number of device manufacturers and major system subcontractors; the third party includes any number of suppliers, subcontractors, and vendors; and in one embodiment, the operator is the owner of the device or the fleet of the device, the administrator responsible for the device or the fleet of the device, the user who operates the device, a leasing company, a military entity, a service organization, etc.

[0037] Now refer to Figure 2 , an apparatus 200 is provided. Figure 2 As shown, an example of the device 200 is a flying device 201, such as an aerospace vehicle, an airplane, air cargo, a flying car, etc. Figure 2 As shown, another example of device 200 is a ground transportation device 202, such as an automobile, truck, heavy equipment, construction equipment, boat, ship, submarine, etc. Figure 2Another example of the device 200 shown is a modular device 203, which includes at least one or more of the following modules: an air module, a payload module, and a ground module. The air module provides air lift or flight capabilities. The payload module provides the ability to transport objects such as cargo or living things (people, animals, etc.). The ground module provides ground mobility capabilities. The solutions disclosed herein are applied to each module individually or to groups of modules, such as air and payload modules, or payload and ground, etc., or all modules.

[0038] Now refer to Figure 3 , shows a more detailed diagram of a flying device 201 implementing an embodiment. In this example, the flying device 201 is composed of Figure 1 An aircraft is produced in apparatus manufacturing and service method 100 and includes fuselage 303 having plurality of systems 304 and interior 306. Implementations of plurality of systems 304 include one or more of propulsion system 308, electrical system 310, hydraulic system 312, and environmental system 314. However, other systems are also candidates for inclusion. Although an aerospace example is shown, different advantageous embodiments are applied to other industries, such as the automotive industry, etc.

[0039] Now refer to Figure 4 , a block diagram of a computing device 400 for generating an SDOMP is provided. In some implementations, the computing device 400 includes one or more processors 402, one or more presentation components 404, and a storage area 410. The disclosed implementations related to the computing device 400 are implemented by various computing devices, including personal computers, laptop computers, smart phones, mobile tablet computers, handheld devices, consumer electronics, special computing devices, etc. The disclosed implementations are also implemented in a distributed computing environment, where tasks are performed by remote processing devices linked through a communication network. In addition, although the computing device 400 is described as seemingly a single device, in one embodiment, multiple computing devices work together and share the described device resources. For example, in one embodiment, the storage area 410 is distributed on multiple devices, the provided processors 402 are accommodated on different devices, and so on.

[0040] In one embodiment, the processor 402 includes any number of processing units that read data from various entities such as the storage area 410. Specifically, the processor 402 is programmed to execute computer executable instructions for implementing various aspects of the present disclosure. In one embodiment, the instructions are executed by a processor, multiple processors, or a processor external to the computing device 400 within the computing device 400. In some implementations, the processor 402 is programmed to execute instructions such as those shown in the flowcharts discussed below and shown in the accompanying drawings. In addition, in some implementations, the processor 402 represents an implementation of simulation technology for performing the operations described herein. For example, these operations are performed by a simulated client computing device 400 and / or a digital client computing device 400.

[0041] The presentation component 404 presents data indications to an operator (e.g., the owner of the device or fleet of devices, an administrator responsible for the device or fleet of devices, or a user operating the device) or to another device. In one implementation, the presentation component includes a display device, a speaker, a printing component, a vibration component, etc. Those skilled in the art will understand and appreciate that computer data is presented in a variety of ways: such as visually in a graphical user interface (GUI); audibly through a speaker; wirelessly between computing devices 400; through a wired connection, or in other ways. In one embodiment, the presentation component 404 is not used when the process and operation are fully automated so that human-computer interaction is reduced or not required. For example, in some embodiments, the process of generating and implementing SDOMPs and changing unplanned tasks to planned tasks and implementing planned tasks is fully automated and without user intervention.

[0042] The storage area 410 stores equipment data 412, approved maintenance procedures (AMP) 414, maintenance costs 416, maintenance improvement goals 418 (provided by the operator), historical maintenance data 420, sensor data 422, data mapping components 424, data correlation components 426, analysis components 428, SDOMP 430, data acquisition and pre-processing components 432, and other data and logic 434. The equipment data 412 includes equipment delivery dates, equipment scrapping dates, and operating hours for a particular type of equipment or equipment fleet, and if leased, the equipment's lease return date, liquidation cycle, and lease matters (e.g., what maintenance events must be performed before the equipment is returned to the lessor). In one embodiment, the AMP 414 includes one or more of the following: maintenance plan data (MPD), equipment-specific maintenance tasks, equipment inspection events, labor hours spent, and access requirements during maintenance. In some implementations, the storage area 410 additionally stores a first improved maintenance program, which is a more optimized or improved version of the AMP 414, although improved using the sensor data 422. Maintenance costs 416 include data related to historical costs incurred by operators for multiple maintenance tasks for a device or fleet of devices. More specifically, the historical costs of each event stored in maintenance costs 416 are based on known costs incurred by one or more maintenance providers (e.g., companies that perform tasks related to the maintenance tasks provided in maintenance costs 416), based on device labor, materials, etc. Historical maintenance data 420 includes information describing historical maintenance events that have been performed from AMP 414, including historical scheduled maintenance data and historical unscheduled maintenance data.

[0043] Sensor data 422 includes sensor-based alarm system (SBAS) device health management (AHM) alarm data, which in some implementations includes data about measurable components, such as any one or all of tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit (APU) oil level, engine oil level, etc. In some implementations, SBAS alarm data includes device health management (AHM) alarm data. Sensor data 422 also includes data from other sensors on the device or in communication with the device. In an embodiment, sensor data 422 includes both historical sensor data and current (e.g., real-time) sensor data. Data mapping component 424 includes a mapping of SBAS alarm data to scheduled maintenance task findings and logic for creating the mapping. Data correlation component 426 establishes an association between non-routine findings and previously related maintenance messages. Analysis component 428 is a software statistical analysis tool that identifies the best event requirements for planning maintenance on a device or device fleet. In one embodiment, statistical analysis is implemented using a known or available statistical application (e.g., but not limited to statistical analysis software (SAS)). SDOMP 430 is the final generated product. The data acquisition and pre-processing component 432 and other data and logic 434 are used to process and store data in support of the data mapping component 424 and the analysis component 428. For example, the other data and logic 434 includes the operating history of the device, the end date of the SDOMP 430 (e.g., the lease end date), and data used to determine the net benefit of the adjustment interval (e.g., the use of the maintenance cost 416).

[0044] Figure 5A graph 500 is shown that provides overlapping probability density curves 502 and 504 and a threshold 506 that drives the detection probability and false alarm probability values. Curve 502 shows the sensor measurement output for a situation that should not cause an alarm. Curve 504 shows the sensor measurement output for a situation that should cause an alarm. Unfortunately, neither the sensor nor the interpretation of the measured sensor output is completely correct. As a result, some situations that should not cause an alarm produce sensor measurement outputs that are indistinguishable from the sensor measurement outputs that should exist in other situations that are worthy of an alarm, and therefore curves 502 and 504 overlap. Threshold 506 is set so that it determines whether an alarm is generated. If the sensed condition exceeds threshold 506, a sensor alarm is issued; conversely, if the sensed condition is less than threshold 506, no sensor alarm is issued. This produces four possibilities: Area 512 falls below curve 502 and is less than threshold 506. Therefore, area 512 represents a situation where the condition should not cause an alarm and an alarm is not issued. This is a correct negative, and there is correctly no alarm. Area 522 falls below curve 502 and is greater than threshold 506. Thus, region 522 represents a false positive (e.g., a false alarm), which is a situation where a condition should not result in an alarm but an alarm is issued. This is a false alarm situation. Region 514 falls below curve 504 and is greater than threshold 506. Thus, region 514 represents a true positive, which is a situation where a condition should result in an alarm and an alarm is issued. This is a correct detection situation. Region 524 falls below curve 504 and is less than threshold 506. Thus, region 524 represents a false negative, which is a situation where a condition should result in an alarm but an alarm is not issued. This is also generally identified as a missed detection situation. In one embodiment, threshold 506 is adjusted upward or downward to increase or decrease the probability of detection (P D ) and false alarm probability (P FA ) to find the desired balance (e.g., optimal threshold). D As indicated by region 514, P FA This is represented by area 522. In general, increasing threshold 506 reduces P FA , but to reduce P D For the cost.

[0045] Figure 6A cumulative distribution function (CDF) curve 602 for the cumulative probability of detecting a condition requiring repair or other maintenance action is shown. In one implementation, the CDF curve 602 is obtained by applying a statistical analysis method to historical inspection data (such as a maintenance record data set for scheduled maintenance findings, and a log book data set for unscheduled maintenance). In another implementation, the CDF curve 602 is generated using the Statistical Analysis Scheduled Maintenance Optimization Engine (SASMO) described in U.S. Patent No. 8,117,007. The CFD curve 602 shows the probability of non-routine findings, and some implementations of the CDF curve 602 are obtained by analyzing historical maintenance data of the device and its components. The maintenance action is, for example, an unscheduled repair, and the maintenance task is, for example, an inspection or preventive maintenance. If the non-routine discovery situation is not remedied in a timely manner, the result may be an impending unscheduled maintenance action. The CDF curve 602 is an estimate using historical maintenance data 420 from one or more operators, and represents the likelihood (probability) of detecting a situation requiring repair or other maintenance action at a certain inspection time after a previous inspection did not identify this situation. A given target acceptable risk level is indicated by line 604. Where line 604 intersects CDF curve 602, this represents an inspection interval T 606 in units of time. Inspections prior to interval T 606 are early inspections and are less likely to result in identification of a situation requiring repair or other maintenance action. Inspections after interval T 606 are late (increased) inspections and, in some cases, are more likely to result in identification of a situation requiring repair or other maintenance action. Without the use of sensor data, the inspection interval is typically set based on interval T 606 determined as described above.

[0046] The maintenance optimization / improvement service helps optimize / improve the operator's maintenance plan based on scheduled and unscheduled maintenance data, combined with sensor data 422. SBAS uses real-time sensor data, such as tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit (APU), and engine oil level to provide diagnostic and predictive services. The diagnostic and predictive capabilities in the SBAS are used to convert unscheduled maintenance into scheduled maintenance, while also optimizing or improving certain inspection task intervals. In one embodiment, an inspection task whose discovery is perfectly predicted by the sensor data 422 is performed virtually, while if the sensor data 422 is only a partial predictor of the maintenance task discovery, a virtual and partial inspection of the task is performed. Therefore, the actual partial inspection interval is increased according to the predicted value of the sensor data 422.

[0047] Figure 7The changed inspection intervals associated with the CDF curve 602 are shown, which are achieved by advantageously using the sensor data 422. When the sensor is an ideal predictor of failure or degradation, the optimal inspection interval approaches infinity, because no inspection is required. However, physical limitations mean that the sensor is a partial predictor of failure or degradation. Relative to relying solely on physical inspections, this allows additional risk to be taken, as shown by the new acceptable risk level indicated by line 704 above line 604. Where line 704 intersects the CDF curve 602, a new inspection interval T' (T initial) 706 is indicated in units of time, which occurs later (longer interval) than interval T 606. Line 604 represents the cumulative probability of physical inspections identifying non-routine findings and setting them to an acceptable risk level. Line 704 indicates the cumulative probability that physical inspections, sensor alarms, or both identify non-routine findings at previous inspection intervals T. Because in some implementations, sensor alarms improve the cumulative probability at a given interval, the interval can be increased (extended) to T' while still maintaining the same acceptable risk level.

[0048] Figure 8 Shows how to determine Figure 7 The CDF curve 602 for identifying the cumulative probability of non-routine findings including degradation and failure is used together with the interval T 606 to determine the target acceptable risk level indicated by line 604. The probability P that a sensor alarm predicts a potential failure is determined, for example, by using the historical maintenance data 420. Since there is a probability P that a sensor alarm predicts a potential failure, the special construction for the equivalent risk level R_equivalent enables the risk that will occur at an extended interval in the absence of sensor data to be estimated. In actual operation, the resulting risk with the extended interval and the use of sensor data does not change. R_equivalent is related to the original target acceptable risk level R and the probability P by the following equation:

[0049] Equation 1: R_equivalent = R + (1-R) ​​× P

[0050] In essence, the sensor alert handles the additional (1−R)×P risk, maintaining the effective risk level R_equivalent as the original target acceptable risk level R. The equivalent risk level (R_equivalent) is represented by line 704 .

[0051] Equation 2: P(J∪S)=P(J)+P(S)-P(J∩S)

[0052] Equation 3: P(J∪S)=P(J)+P(S)-P(J|S)×P(S); According to the Bayesian equation

[0053] Equation 4: P(J∪S) = P(J) + P(S) - P(J) × P(S); Since J and S operate independently

[0054] Equation 5: P(J∪S)=P(J)+(1-P(J))×P(S)

[0055] Where P(J∪S) is the probability that J or sensor S, or both, detects a physical inspection that is not a routine finding, P(J) is the probability that J alone detects a physical inspection that is not a routine finding, and P(S) is the probability that sensor S alone detects a non-routine finding (correct detection). The point where line 704 intersects CDF curve 602 represents the new inspection interval T'706 in units of time. Undetected failure probability distribution curve 812 represents the likelihood of a failure that was not predicted by the sensor alarm. The original target acceptable risk level indicated by line 604 intersects the undetected failure probability distribution curve 812 at a location corresponding to the new inspection interval T'706, showing how the sensor allows the inspection interval to be increased (from T 606 to T'706) while maintaining the original target acceptable risk level. In some implementations, if the probability of sensor detection is random, with a known distribution (e.g., a Gaussian distribution), a convolution method is used to estimate the risk reduced by using sensor data 422.

[0056] As an illustrative example, a maintenance task includes: "General visual inspection of engine oil filter element bypass condition" with an interval T of 150 flight hours. The initial interval T of 150 hours is derived from the CDF curve 602 with an acceptable risk level R of 0.4. The sensor continuously measures the oil filter element bypass condition. If the oil filter element bypass exceeds a certain threshold, the sensor generates an alarm from the sensor. Two scenarios are presented:

[0057] Case 1: Ideal sensor with P(S) = 1. If there is a fault, the sensor will capture it with 100% certainty. In this case, Equation 1 becomes:

[0058] Equation 6: R_equivalent = R + (1-R) ​​× P (S) = 0.4 + 0.6 × 1

[0059] New physical inspection intervals are pushed out indefinitely. That is, the inspection task is removed from the physical inspection, and only virtual inspection is required.

[0060] Case 2: Imperfect sensor with P(S) = 0.8. If there is a fault, the sensor will capture it with 80% certainty. In this case, Equation 1 becomes:

[0061] Equation 7: R_equivalent = R + (1-R) ​​× P (S) = 0.4 + 0.6 × 0.8 = 0.88

[0062] New physical inspection intervals increase, although not indefinitely.

[0063] Fig. 9 4 is a block diagram of an implementation of a process flow 900 for implementing aspects of the present disclosure. Input data includes: AMP 414 and maintenance improvement goals 418, which together specify tasks, intervals, and acceptable risks; SBAS data and alerts 902, which are part of sensor data 422; fault isolation information 904; historical scheduled maintenance data 906; and historical unscheduled maintenance data 908. The historical scheduled maintenance data 906 and the historical unscheduled maintenance data 908 are from the historical maintenance data 420. The data acquisition and preprocessing phase includes using the data acquisition and preprocessing component 432 to acquire and / or process the SBAS data and alerts 902 with processing 932a, the fault isolation information 904 with processing 932b, and the historical scheduled maintenance data 906 and the historical unscheduled maintenance data 908 with processing 932c. These results are fed to the data mapping component 424. The task and alert selection process 910 selects various maintenance tasks as candidates for removal from physical inspections and / or additions. Each task is then fed into the analysis component 428. The analysis component 428 then makes a determination at decision operation 912 whether to maintain the original interval at 914, remove the task from the physical inspection at 916 (making it a virtual inspection), or increase the interval at 918. In some implementations, the analysis component 428 is configured to generate the data for the undetected failure probability distribution curve 812.

[0064] The decision operation 912 uses the predicted values ​​of the sensor data 422 and the AMP 414 and the maintenance improvement goal 418. If the sensor data 422 has no predicted value, i.e., the failure and non-routine findings cannot be predicted by any sensor data 422 or alarm, the original AMP interval is retained. If the sensor data 422 has complete predicted value, i.e., the failure and non-routine findings are reliably predicted by the sensor data 422 or alarm, the maintenance tasks such as inspections are removed from the physical inspection. However, if the sensor data 422 has partial predicted value, i.e., the failure and non-routine findings can be predicted with reasonable probability by the sensor data 422 or alarm, the maintenance tasks are escalated according to the reliability of the detection. In this way, the SDOMP 430 is generated based on at least the predicted values ​​of the sensor data 422 and the maintenance improvement goal 418. In some implementations, the SDOMP 430 includes intervals that are adjusted (e.g., increased or decreased), and also reflects the removed tasks and some unchanged AMP intervals.

[0065] Fig.101021. The present invention provides an implementation of SBAS data and alert conditions from different sensors (e.g., sensors produced by different component vendors) stored in SBAS data and alerts 902 having different formats. For example, alert data 1011 and 1012 from sensors 1001 and 1002, respectively, are in a first format 1021. Alert data 1013 and 1014 from sensors 1003 and 1004, respectively, are in a second, different format 1022. This illustrates the need for data acquisition and pre-processing component 432 to be able to receive data from different sources in different formats and process it as needed for downstream consumption. SBAS data and alerts 902 includes sensors from different component vendors and from different sources and therefore in inconsistent formats.

[0066] Fig.11 1100 is a block diagram of an implementation of a sensor mapping operation. Operation 1102 is part of the task and alert selection process 910, in which candidate SBAS alerts are identified. In some implementations, operation 1102 includes adjusting an alert threshold (e.g., Figure 5 The threshold 506 of the alarm is used to find the optimal alarm threshold. If the sensor alarm provides too many false alarms to be relied upon, the resulting unscheduled inspections will increase maintenance costs. In some examples, the sensor itself is adjusted to change the alarm condition to a lower false alarm rate. This is used in conjunction with the fault isolation information 904, which in some implementations is derived from a fault isolation manual provided by the component manufacturer. These data sets are then analyzed with recommended maintenance actions 1104.

[0067] Additionally, in some implementations, maintenance tasks 1106 derived from historical maintenance data 420 are used to identify unscheduled maintenance task information. Unscheduled maintenance task candidates 1108 are selected, which provide failure information for various components and data about non-routine findings. In some implementations, data about unscheduled maintenance task candidates 1108 is also derived from historical maintenance data 420. Then, at 1112, maintenance actions 1110 are compared to recommended maintenance actions 1104 to map sensor data 422 to maintenance findings. In some implementations, the logical mapping of sensor data (including sensor alarms) to maintenance actions is followed by another stage of using historical data to determine whether the number of false alarms is too high.

[0068] Fig.1212 is an illustration of a timeline 1200 of related events associated with generating an SDOMP 430: a first event 1202 and a second event 1204. The first event 1202 is a maintenance message, such as a sensor alarm or other reported sensor data, followed by the second event 1204 after an interval 1206. The second event 1204 is a non-routine finding, identified as non-routine finding X. Initially, no correlation may be identified between the first event 1202 and the second event 1204. However, during the mapping operation, a correlation is identified between the first event 1202 and the second event 1204. For example, if a large number of non-routine findings have been preceded by a related maintenance message, a correlation can often be established.

[0069] Fig.13 The relationship between maintenance tasks (via non-routine discovery) and maintenance messages in a conceptual temporal relationship network 1300 is shown. Fig.13 Provides a basis for understanding Fig.11 Additional perspectives of block diagram 1100 of FIG. 40. In some examples, sensor data 422 (whether alarms or general sensor data) is not mapped one-to-one with maintenance tasks (such as inspections) and / or non-routine findings. In embodiments, there are overlapping one-to-many relationships, many-to-one relationships, and combinations. For example, multiple sensor alarms are associated with a first non-routine finding, and one or more of these multiple sensor alarms are additionally associated with multiple other non-routine findings.

[0070] Typically, sensor data 422 brings up two considerations: probability of detection (accompanied by probability of false alarm) and relevance. The first consideration, probability of detection, is a reliability consideration that addresses the question of how reliably sensor data 422 predicts non-routine findings. It is often possible to establish correlations with statistical analysis that calculates correlations. Correlation is a mathematical specification of the correlation between two events. One correlation method used in some implementations is the Pearson correlation coefficient (PCC), which is a measure of the linear correlation between two variables X and Y. Perfect correlation means that two events always occur together, while for unrelated events, when one event occurs, there is no probative value in predicting whether the second event will occur. Analyze the reliability and relevance of sensor data 422 to maintenance tasks, non-routine findings, and maintenance actions.

[0071] The conceptual temporal relationship network 1300 addresses questions such as: What is the relevance of a particular sensor or group of sensors? How reliable is the sensor data 422? What role does the sensor data 422 provide to improve the predictive value? For example, the importance of the maintenance message 1302 is identified using the fault isolation information 904. The maintenance tasks 1106 are used to identify non-routine findings 1308. Relationships are then found between the fault isolation information 904 and the maintenance actions 1110, and relationships are found between the non-routine findings 1308 and the maintenance actions 1110. This then completes the mapping between the maintenance messages 1302 and the maintenance tasks 1106 to facilitate quantifying the relevance of the maintenance findings with the associated sensor alarms and other sensor data.

[0072] FIG. 14A to FIG. 14D Shows the net benefit of various scenarios for utilizing sensor data. Figure 8 , describes a set of example scenarios where the physical inspection interval is T. T is the inverse function of the risk level R (plotted as CDF curve 602), and T' is the increased interval. Therefore, for the original and increased intervals, the number of scheduled maintenance actions per unit time is i1 / T and 1 / T', respectively. The increased benefit B is calculated as follows:

[0073] Equation 8: B = (1 / T-1 / T') × C_periodic

[0074] Where C_Periodically is the cost of periodic maintenance of the task.

[0075] Fig.14A Bar graph 1400 showing Scenario 1: The airline has a process for handling sensor-based alerts. Therefore, there is already a business process in place to handle inspections triggered by alerts, so inspection costs are not included. As shown in Equation 8, the net benefit of the SBAS addition is B. In bar graph 1400, bar 1402 shows the original cost of scheduled maintenance with an interval of T. Bar 1404 shows the new cost per unit time for scheduled maintenance with an interval of T', and benefit B is shown as difference 1408.

[0076] Fig. 14B Bar graph 1410 shows scenario 2.1: the airline has no process for handling sensor-based alerts, and sensor thresholds cannot be changed to reduce false alarms. As shown in Equation 9, the net benefit of SBAS addition is B:

[0077] Equation 9: B = (1 / T - 1 / T') × C_periodic - P FA ×C_Irregular

[0078] Where P FAis the probability of a false sensor alarm, and C_unscheduled is the cost of unscheduled maintenance for the task. In bar graph 1410, bar 1416 shows the cost of unscheduled maintenance at intervals of T', so it is added on top of bar 1404. The net benefit B is shown as difference 1418. If the sensor false alarms are so many that the net benefit is negative (e.g., more money is spent), then the sensor data should not be used to increase the maintenance interval.

[0079] Fig. 14C The bar graph 1420 of scenario 2.2 is shown, except that the sensor alarm thresholds can be changed (e.g., Figure 5 threshold 506) to reduce P FA Otherwise, it is similar to Case 2.1. Initially, P FA and C_ irregularly generate irregular maintenance costs, as shown by bar 1426 on top of bar 1404, i.e., the costs are so high that the net benefit is negative. Therefore, the sensor alarm threshold is adjusted to reduce P FA . This has the effect of lowering the probability of detection P(S), so the new interval will be T″ with a lower P(S).

[0080] Fig.14D The bar graph 1430 for scenario 2.2 is also shown, but reflects the new interval T″. In the bar graph 1430, bar 1434 shows the new cost of scheduled maintenance with interval T″, which is higher than (reflecting the higher cost) FIG. 14A to FIG. 14C Bar 1404 in FIG. Bar 1436 on top of bar 1434 shows the P due to irregular and reduced C_. FA The net benefit B is shown as difference 1438. Bar graph 1430 shows a situation where there is an alarm threshold, where P FA and P D The combination of (the probability of correct detection and alarm) produces a net benefit greater than zero. In summary, FIG. 14A to FIG. 14D It is shown that, although reliance on sensor data 422 reduces maintenance events when the probability of false alarms is low enough, reliance on sensor data 422 unnecessarily increases maintenance events, thereby potentially increasing operator costs, when the probability of false alarms is too high.

[0081] Fig.15A flowchart 1500 is provided that illustrates an implementation of an operation for determining whether to increase a maintenance task. Flowchart 1500 provides further explanation of decision operation 912. Selected historical sensor data 422a, which is a subset of sensor data 422, is combined with historical maintenance data 420 in a correlation operation 1502, which, in some implementations, is represented by a conceptual temporal relationship network 1300. Decision operation 1504 determines whether the sensor data 422 can predict a failure or other non-routine finding. If not, the result is a decision 914 to retain the original maintenance interval. It should be noted that the determination in decision operation 1504 is unlikely to be a clear yes or no result. In one example, the determination is made based on a predicted value calculation that rarely reaches 0.0, but has a low value for a no result. That is, in some implementations, the predicted value is compared to a threshold value to provide a yes or no result.

[0082] If the sensor data 422 has predicted a failure or non-routine finding, decision operation 1506 determines whether the sensor data 422 has fully predicted the non-routine finding (e.g., degradation or failure or other non-routine finding). If so, the result is a decision 916 to remove the maintenance task from the physical inspection and use a virtual inspection (via the sensor). Similar to decision operation 1504, some implementations of decision operation 1506 use a second threshold value without requiring an absolutely perfect sensor data prediction value. If the sensor data 422 is not fully predictive, the sensor data 422 may have only partially predicted the failure or other non-routine finding (reducing rather than eliminating the risk), and the result is a decision 918 to increase the maintenance interval. Part of this process (increasing the maintenance interval) is Figure 8 Modification of the detection probability is shown. Referring to the threshold used in decision operation 1504, when the predicted value of the sensor data 422 meets the threshold, the inspection interval is increased.

[0083] Now refer to Fig.16 , flowchart 1600 illustrates an implementation of operations for generating SDOMP 430. In one embodiment, Fig.16 The illustrated operations are performed by the analysis component 428 providing instructions to the one or more processors 402. Operation 1602 includes receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed. In some implementations, the first maintenance plan includes an AMP (e.g., AMP 414). In some implementations, the device type is an aircraft (e.g., the flying device 201).

[0084] Operation 1604 includes receiving historical sensor data (e.g., from sensor data 422); operation 1606 includes receiving historical scheduled maintenance data (e.g., from historical maintenance data 420), and operation 1608 includes receiving historical unscheduled maintenance data (e.g., from historical maintenance data 420). In some implementations, the received sensor data includes SBAS alert data. In some implementations, the SBAS alert data includes at least one data type selected from the group consisting of: tire pressure, oxygen pressure, hydraulic fluid, APU oil level, and engine oil level.

[0085] Operation 1610 includes correlating historical sensor data with historical scheduled maintenance data and historical unscheduled maintenance data to determine a predicted value for sensor data 422. In some implementations, determining the predicted value for sensor data 422 includes determining a probability that a sensor alarm corresponds to a non-routine finding. Operation 1612 includes receiving an operational history of the device, and operation 1614 includes determining a net benefit of adjusting the interval. False alarms (false positives) and missed detections (false negatives) increase maintenance costs. See, e.g., Fig. 14C and Fig.14D In some implementations, the cost information is weighted according to the probability of false alarms and missed detections. For example, when a first maintenance task is postponed due to reliance on a sensor, and a more expensive second maintenance task is required as a result, then the cost of the more expensive second maintenance task is used in the determined cost. Operations 1612 and 1614 use data from other data and logic 434 and in combination with Figure 17 to Figure 18 Further description.

[0086] Operation 1616 includes constructing the CDF curve 602, and generating an SDOMP (e.g., SDOMP 430) based on at least the predicted value of the sensor data 422, the generated SDOMP 430 including the adjusted interval in the defined interval. In some implementations, generating the SDOMP 430 based on at least the predicted value of the sensor data 422 and the maintenance improvement goal 418 includes increasing the inspection interval when the predicted value of the sensor data 422 meets the threshold. In some implementations, generating the SDOMP 430 also includes generating the SDOMP 430 based on at least the operation history of the device. In some implementations, generating the SDOMP 430 also includes generating the SDOMP 430 based on at least the end date. In some implementations, generating the SDOMP also includes generating the SDOMP 430 based on at least the cost associated with the maintenance task. Operation 1618 then includes presenting the generated SDOMP 430, for example, on the presentation component 404.

[0087] Fig.1717 is a block diagram illustrating a process 1700 in which the analysis component 428 generates the SDOMP 430. The analysis component 428 uses the device data 412 and the AMP 414 in operation 1602; the historical maintenance data 420 is used for operations 1606 and 1608. In operation 1604, the analysis component 428 uses the sensor data 422 (e.g., historical sensor data); and uses the results of operations 1604, 1606, and 1608 in operation 1610. The analysis component 428 uses other data and logic 434 in operations 1612 and 1614. In some implementations, the other data and logic 434 includes the operation history of the device, the end date of the SDOMP 430 (e.g., the lease end date), and data used to determine the net benefit of the adjustment interval (e.g., the use maintenance cost 416). The results of operations 1602, 1610, 1612, and 1614 are used in operation 1618 to generate the SDOMP 430.

[0088] Fig.18 1 is a block diagram of a system 1800 for generating SDOMP. The system 1800 is one implementation of a suitable computing environment and is not intended to suggest any limitation on the scope of use or functionality of the invention. The system 1800 should also not be interpreted as having any dependency or requirement on any one or combination of components / modules illustrated.

[0089] The implementations and embodiments disclosed herein are described in the general context of computer code or machine-usable instructions, including computer-executable instructions such as program components executed by a computer or other machine such as a personal data assistant or other handheld device. In general, program components, including routines, programs, objects, components, data structures, refer to code that performs a specific task or implements a specific abstract data type. The disclosed implementations are practiced in various system configurations, including personal computers, laptop computers, smart phones, mobile tablet computers, handheld devices, consumer electronics, special-purpose computing devices, etc. The disclosed implementations are also implemented in distributed computing environments, where tasks are performed by remote processing devices linked through a communication network.

[0090] System 1800 includes a computing device (e.g., computing device 400) communicatively coupled to a network 1818. Computing device 400 includes a bus 1816 that directly or indirectly couples the following devices: storage area 410, one or more processors 402, one or more presentation components 404, input / output (I / O) ports 1808, I / O components 1810, power supply 1812, and network components 1814. System 1800 should not be interpreted as having any dependency or requirement on any single component or combination of components shown therein. Although system 1800 is described as appearing to be a single device, in one embodiment, multiple computing devices work together and share the described device resources. For example, in one embodiment, storage area 410 is distributed across multiple devices, processors 402 provided are housed on different devices, and the like.

[0091] Bus 1816 represents one or more buses (e.g., an address bus, a data bus, or a combination thereof). Although shown with lines for clarity, Fig.18 , but in reality, it is not very clear to depict the individual components, and metaphorically, the lines would more accurately be gray and fuzzy. For example, one would think of a presentation component such as a display device as an I / O component. Also, a processor has memory. This is the nature of the art, and Fig.18 The illustrations are merely examples of systems or computing devices that may be used in conjunction with one or more embodiments of the present invention. No distinction is made between categories such as "workstation," "server," "laptop," "handheld device," etc., as all categories are considered to be common in the Fig.18 , and herein refers to a “computing device”.

[0092] In one embodiment, the storage area 410 includes any computer readable medium discussed herein. In one embodiment, the storage area 410 is used to store and access instructions configured to perform various operations disclosed herein. In some implementations, the storage area 410 includes computer storage media in the form of volatile and / or non-volatile memory, removable or non-removable memory, a data disk in a virtual environment, or a combination thereof.

[0093] In one embodiment, the processor 402 includes any number of processing units that read data from various entities such as storage area 410 or I / O component 1810. Specifically, the processor 402 is programmed to execute computer executable instructions for implementing various aspects of the present disclosure. In one embodiment, the instructions are executed by a processor, multiple processors, or a processor external to the computing device 400 within the computing device 400. In some implementations, the processor 402 is programmed to execute instructions such as those shown in the flowcharts discussed below and shown in the accompanying drawings. In addition, in some implementations, the processor 402 represents an implementation of simulation technology for performing the operations described herein. For example, these operations are performed by simulated client computing devices and / or digital client computing devices.

[0094] The presentation component 404 presents data indications to an operator (e.g., an owner of a device or fleet of devices, an administrator responsible for a device or fleet of devices, or a user operating a device) or to another device. In one implementation, the presentation component includes a display device, a speaker, a printing component, a vibration component, etc. Those skilled in the art will understand and appreciate that computer data is presented in a variety of ways, such as visually in a graphical user interface (GUI); audibly through a speaker; wirelessly between computing devices 400; through a wired connection, or in other ways.

[0095] Port 1808 allows computing device 400 to be logically coupled to other devices, some of which are built-in, including I / O components 1810. Implementations of I / O components 1810 include, but are not limited to, for example, a microphone, keyboard, mouse, joystick, game controller, satellite dish, scanner, printer, wireless device, etc.

[0096] In some implementations, the network component 1814 includes a network interface card and / or computer executable instructions (e.g., a driver) for operating a network interface card. In one embodiment, communication between the computing device 400 and other devices occurs over any wired or wireless connection using any protocol or mechanism. In some implementations, the network component 1814 is operable to use a transport protocol over public, private, or hybrid (public and private), over short-range communication technologies (e.g., near field communication (NFC), brand communications, etc.) or a combination thereof.

[0097] Although described in conjunction with computing device 400, implementations of the present disclosure can be implemented with many other general or special computing system environments, configurations, or devices. Implementations of known computing systems, environments, and / or configurations suitable for various aspects of the present disclosure include, but are not limited to, smartphones, mobile tablet computers, mobile computing devices, personal computers, server computers, handheld or laptop devices, multiprocessor systems, game consoles, microprocessor-based systems, set-top boxes, programmable consumer electronics, mobile phones, mobile computing and / or communication devices in wearable or accessory form factors (e.g., watches, glasses, headphones or earphones), network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, VR devices, holographic devices, etc. Such systems or devices accept input from a user in any manner, including from input devices such as a keyboard or pointing device, via gesture input, proximity input (such as by hovering), and / or via voice input.

[0098] In one embodiment, implementations of the present disclosure are described in the general context of computer executable instructions executed by one or more computers or other devices in software, firmware, hardware, or a combination thereof, such as program modules. In one embodiment, computer executable instructions are organized into one or more computer executable components or modules. In general, program modules include, but are not limited to, routines, programs, objects, components, and data structures that perform specific tasks or implement specific abstract data types. In one embodiment, aspects of the present disclosure are implemented with such components or modules of any number and organization. For example, aspects of the present disclosure are not limited to specific computer executable instructions or specific components or modules shown in the accompanying drawings and described herein. Other implementations of the present disclosure include different computer executable instructions or components with more or less functions than those shown and described herein. In implementations involving general-purpose computers, aspects of the present disclosure convert general-purpose computers into special-purpose computing devices when configured to execute instructions described herein.

[0099] As an example and not limitation, computer-readable media include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable memory implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, etc. Computer storage media are tangible and mutually exclusive with communication media. Computer storage media are implemented in hardware and exclude carrier waves and propagation signals. Computer storage media for the purposes of this disclosure are not signals themselves. In one implementation, computer storage media include hard disks, flash drives, solid-state memory, phase change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, cassettes, tapes, disk storage or other magnetic storage devices, or any other non-transmission medium for storing information accessed by a computing device. In contrast, communication media typically embodies computer readable instructions, data structures, program modules, etc. in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.

[0100] The following paragraphs describe other aspects of the present disclosure:

[0101] A1. A method for generating SDOMP, the method comprising:

[0102] receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed;

[0103] Receive historical sensor data, historical regular maintenance data, and historical irregular maintenance data;

[0104] Correlating the historical sensor data with the historical scheduled maintenance data and the historical unscheduled maintenance data to determine a predicted value of the sensor data;

[0105] receiving one or more maintenance improvement goals for an operator of a device of the device type; and

[0106] Based at least on the predicted values ​​of the sensor data and the maintenance improvement goal, an SDOMP is generated, the generated SDOMP including adjusted intervals in the defined intervals.

[0107] A2. The method of A1, wherein the first maintenance plan comprises an approved maintenance plan.

[0108] A3. The method of A1 or A2, wherein the device type is an aircraft.

[0109] A4. A method according to any of A1-A3, wherein the sensor data includes aircraft health management (AHM) alert data, and wherein the AHM alert data includes at least one data type selected from the following items, the following items including:

[0110] Tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit (APU) oil level and engine oil level.

[0111] A5. A method according to any of A1-A4, wherein determining the predicted value of the sensor data includes determining a probability that the sensor alert corresponds to a non-routine finding.

[0112] A6. A method according to any of A1-A5, wherein one or more maintenance improvement goals of the operator include an acceptable risk level.

[0113] A7. A method according to any one of A1-A6, wherein generating the SDOMP based at least on the predicted value of the sensor data and the maintenance improvement target includes: increasing the inspection interval when the predicted value of the sensor data meets a threshold.

[0114] A8. The method according to any one of A1-A7, further comprising:

[0115] The operating history of the receiving device, and

[0116] Wherein generating the SDOMP further comprises generating the SDOMP based at least on an operation history of the device.

[0117] A9. The method according to any one of A1-A8, further comprising:

[0118] Determine the end date of the SDOMP, and

[0119] Wherein generating the SDOMP further comprises generating the SDOMP based on at least an end date.

[0120] A10. The method according to any one of A1-A9, further comprising:

[0121] Determine the net benefit of adjusting the interval, and

[0122] Wherein generating the SDOMP further comprises generating the SDOMP based at least on a cost associated with the maintenance task.

[0123] A11. A system for generating SDOMP, the system comprising:

[0124] one or more processors; and

[0125] A storage area stores an analysis component, which, when executed by one or more processors, causes the one or more processors to perform the following operations, the operations comprising:

[0126] receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed;

[0127] Receive historical sensor data, historical regular maintenance data, and historical irregular maintenance data;

[0128] correlating the historical sensor data with the historical scheduled maintenance data and the historical unscheduled maintenance data to determine a predicted value of the sensor data;

[0129] receiving one or more maintenance improvement goals for an operator of a device of the device type; and

[0130] Based at least on the predicted values ​​of the sensor data and the maintenance improvement goal, an SDOMP is generated, the generated SDOMP including adjusted intervals in the defined intervals.

[0131] A12. The system of A11, wherein the first maintenance plan comprises an approved maintenance plan.

[0132] A13. A system according to A11 or A12, wherein the device type is an aircraft.

[0133] A14. A system according to any of A11-A13, wherein the sensor data includes AHM alarm data, and wherein the AHM alarm data includes at least one data type selected from the following items, the following items including:

[0134] Tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit (APU) oil level and engine oil level.

[0135] A15. A system according to any of A11-A14, wherein determining the predicted value of the sensor data includes determining a probability that the sensor alert corresponds to a non-routine finding.

[0136] A16. A system according to any of A11-A15, wherein one or more maintenance improvement goals of the operator include an acceptable risk level.

[0137] A17. A system according to any one of A11-A16, wherein generating the SDOMP based on at least the predicted value of the sensor data and the maintenance improvement target includes: gradually increasing the inspection interval when the predicted value of the sensor data meets a threshold.

[0138] A18. A system according to any one of A11-A17, wherein the operation further comprises:

[0139] The operating history of the receiving device, and

[0140] Wherein generating the SDOMP further comprises generating the SDOMP based at least on an operation history of the device.

[0141] A19. A system according to any one of A11-A18, wherein the operation further comprises:

[0142] Determine the end date of the SDOMP, and

[0143] Wherein generating the SDOMP further comprises generating the SDOMP based on at least an end date.

[0144] A20. A system according to any one of A11-A19, wherein the operation further comprises:

[0145] Determine the net benefit of adjusting the interval, and

[0146] Wherein generating the SDOMP further comprises generating the SDOMP based at least on a cost associated with the maintenance task.

[0147] When introducing elements of aspects of the present disclosure or implementations thereof, the articles "a," "an," "the," and "said" are intended to mean that there are one or more of the elements. The terms "comprising," "including," and "having" are intended to be inclusive and mean that there may be additional elements in addition to the listed elements. The term "implementation" is intended to mean "an example of ..." The phrase "one or more of: A, B, and C" means "at least one of A and / or at least one of B and / or at least one of C."

[0148] Having described various aspects of the disclosure in detail, it will be apparent that modifications and variations are possible without departing from the scope of the various aspects of the invention as defined in the appended claims. As various changes may be made to the above-described constructions, products, and methods without departing from the scope of the various aspects of the disclosure, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not restrictive.

Claims

1. A method for generating a sensor-driven optimized maintenance program, the method comprising: receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed; Receive historical sensor data, historical regular maintenance data, and historical irregular maintenance data; associating the historical sensor data with the historical scheduled maintenance data and the historical unscheduled maintenance data to determine a predicted value of the sensor data; receiving one or more maintenance improvement goals for an operator of a device of the device type; as well as removing the maintenance task from the first maintenance plan to generate the sensor-driven optimized maintenance program based on the predicted value satisfying a threshold, and replacing a physical inspection associated with the maintenance task with a virtual inspection performed by a sensor; generating the sensor-driven optimized maintenance program based at least on the predicted value failing to satisfy the threshold and the maintenance improvement goal, the generated sensor-driven optimized maintenance program comprising an adjusted interval, the adjusted interval being delayed in time compared to the interval; as well as The sensor-driven optimized maintenance program generated based on execution causes the maintenance tasks at the adjusted intervals to be performed at delayed times.

2. The method according to claim 1, wherein: The first maintenance plan comprises an approved maintenance plan.

3. The method according to claim 1 or 2, wherein: The device type is an aircraft.

4. The method according to claim 1 or 2, wherein: The sensor data includes device health management alert data, and wherein the device health management alert data includes at least one data type selected from the list consisting of: Tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit oil level and engine oil level.

5. The method according to claim 1 or 2, wherein: Determining the predicted value of the sensor data includes determining a probability that a sensor alert corresponds to a non-routine finding.

6. The method according to claim 1 or 2, wherein: The one or more maintenance improvement goals for the operator include an acceptable level of risk.

7. The method according to claim 1 or 2, wherein: Generating the sensor-driven optimized maintenance program based at least on the predicted value of the sensor data and the maintenance improvement goal includes increasing an inspection interval when the predicted value of the sensor data meets a threshold.

8. The method according to claim 1 or 2, further comprising: receiving an operation history of the device, and Wherein generating the sensor-driven optimized maintenance program further comprises generating the sensor-driven optimized maintenance program based at least on the operation history of the device.

9. The method according to claim 1 or 2, further comprising: determining an end date for the sensor-driven optimized maintenance program, and Wherein generating the sensor-driven optimized maintenance program further comprises generating the sensor-driven optimized maintenance program based at least on the end date.

10. The method according to claim 1 or 2, further comprising: Determine the net benefit of adjusting the interval, and Wherein generating the sensor driven optimized maintenance program further comprises generating the sensor driven optimized maintenance program based at least on a cost associated with the maintenance task.

11. A system for generating a sensor-driven optimized maintenance program, the system comprising: one or more processors; as well as A storage area stores an analysis component, and when the analysis component is executed by the one or more processors, the one or more processors are caused to perform operations, the operations comprising: receiving a first maintenance plan for a device type, the first maintenance plan defining maintenance tasks for the device type and defining intervals at which the maintenance tasks are to be performed; Receive historical sensor data, historical regular maintenance data, and historical irregular maintenance data; associating the historical sensor data with the historical scheduled maintenance data and the historical unscheduled maintenance data to determine a predicted value of the sensor data; receiving one or more maintenance improvement goals for an operator of equipment of the equipment type; and removing the maintenance task from the first maintenance plan to generate the sensor-driven optimized maintenance program based on the predicted value satisfying a threshold, and replacing a physical inspection associated with the maintenance task with a virtual inspection performed by a sensor; generating the sensor-driven optimized maintenance program based at least on the predicted value failing to satisfy the threshold and the maintenance improvement goal, the generated sensor-driven optimized maintenance program comprising an adjusted interval, the adjusted interval being delayed in time compared to the interval; and Based on executing the generated sensor-driven optimized maintenance program, the maintenance tasks at the adjusted intervals are performed at delayed times.

12. The system according to claim 11, wherein: The first maintenance plan comprises an approved maintenance plan.

13. The system according to claim 11 or 12, wherein: The device type is an aircraft.

14. The system according to claim 11 or 12, wherein: The sensor data includes device health management alert data, and wherein the device health management alert data includes at least one data type selected from the list consisting of: Tire pressure, oxygen pressure, hydraulic fluid, auxiliary power unit oil level and engine oil level.

15. The system according to claim 11 or 12, wherein: Determining the predicted value of the sensor data includes determining a probability that a sensor alert corresponds to a non-routine finding.

16. The system according to claim 11 or 12, wherein: The one or more maintenance improvement goals for the operator include an acceptable level of risk.

17. The system according to claim 11 or 12, wherein: Generating the sensor-driven optimized maintenance program based at least on the predicted value of the sensor data and the maintenance improvement goal includes increasing an inspection interval when the predicted value of the sensor data meets a threshold.

18. The system according to claim 11 or 12, wherein: The operations also include: receiving an operation history of the device, and Wherein generating the sensor-driven optimized maintenance program further comprises generating the sensor-driven optimized maintenance program based at least on the operation history of the device.

19. The system according to claim 11 or 12, wherein: The operations also include: determining an end date for the sensor-driven optimized maintenance program, and Wherein generating the sensor-driven optimized maintenance program further comprises generating the sensor-driven optimized maintenance program based at least on the end date.

20. The system according to claim 11 or 12, wherein: The operations also include: Determine the net benefit of adjusting the interval, and Wherein generating the sensor driven optimized maintenance program further comprises generating the sensor driven optimized maintenance program based at least on a cost associated with the maintenance task.

Images