Systems and methods for scalable automated maintenance optimization
Patent Information
- Application Number
- JP2023527985
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-10-14
- Filing Date
- 2021-11-03
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2041-11-03
AI Technical Summary
Capital-intensive industries face challenges in determining the age of parts across a portfolio of assets due to inconsistent and incomplete maintenance data, leading to inaccurate predictive maintenance strategies and inefficient resource allocation, especially when dealing with complex machinery in remote locations.
A scalable automated maintenance optimization system that integrates maintenance records using natural language processing and automated statistical analysis to calculate the age of parts, determine failure patterns, and optimize replacement strategies, incorporating data from work orders, sensor measurements, and asset ontologies.
Enables accurate and efficient calculation of part age and failure patterns, allowing for optimized maintenance schedules and inventory management across a portfolio of assets, reducing downtime and costs while improving reliability and resource utilization.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
Cross-reference
[0003] , , ,
[0001] This application claims priority to U.S. Patent Application No. 17 / 484,022, filed September 24, 2021, which is a continuation-in-part of U.S. Patent Application No. 17 / 165,045, filed February 2, 2021, which claims priority to U.S. Provisional Patent Application No. 63 / 110,428, filed November 6, 2020. This application also claims priority to U.S. Provisional Patent Application No. 63 / 255,546, filed October 14, 2021. The entire disclosure of each of the documents cited in this section is hereby incorporated by reference in its entirety into this specification.
Technical Field
[0002] Aspects of the present disclosure generally relate to systems and methods for improving maintenance strategies applied to portfolios of assets in capital-intensive industries using transaction data from maintenance management systems, as well as applications of data science that encapsulate and automate the expertise of experts in the field of maintenance. More particularly, without limitation, the present invention relates to an accurate determination of the age of parts across a portfolio of assets based on extracting and integrating maintenance records using natural language processing, asset ontologies, and automated statistical analysis.
Background Art
[0003] For example, without limitation, capital-intensive industries such as mining, freight railroads, shipping, power generation, and oil and gas operate complex machinery 24 hours a day under difficult remote conditions. Operations may be limited to a single site or may span geographical regions, continents, or multiple sites on the earth. Site operations, including asset management and parts replacement, can be managed locally at each site. However, it is more common and useful for data to be interchangeable between different sites of the same type (e.g., mines in different locations) such that the local operations at each site can view the operations and maintenance data of other sites to benchmark and modify the local operations.
[0004] When a piece of equipment is replaced as part of planned or unplanned maintenance, the replacement work is requested in a notification. The task description is recorded in the relevant work order, which may be fed into the operations maintenance management system and stored locally. The work order may contain data included in the notification. The information from the work order informs future parts replacement strategies to maximize uptime cost-effectively. Each local site's operations maintenance management system may feed transaction data to a central server acting as a client to the global operations maintenance management system.
[0005] In typical capital-intensive organizations, reliability engineers are tasked with extracting and analyzing data from the operations maintenance management system to determine the optimal maintenance strategy. The analysis generally involves focusing on a specific part at a time, identified by its type and location within the asset, and collecting all work order records for when that part was replaced to calculate overall failure statistics. This analysis becomes even more complex when multiple parts of a single part type are installed on an asset, and only some of those parts are replaced. Furthermore, work orders are often incomplete, lacking crucial information to identify the location of the part being replaced or the cause of the replacement, which could otherwise identify its location.
[0006] These discrepancies and inconsistencies can lead to work instructions being completed prematurely and inaccurately. This can occur due to operational pressures. Furthermore, functionally equivalent parts may be sourced from different manufacturers and therefore identified by different part number identification codes. These challenges can make it take months to manually reconstruct the history of a particular part type that fits into an asset portfolio. An asset portfolio can include fixed factory assets and fleets of similar assets. Even if the history of a part type can be manually reconstructed, inconsistent and incomplete data can prevent accurate determination of the age of each instance of that part at the time of replacement across the asset portfolio for a particular operation.
[0007] Determining the age is crucial for generating predictive parts maintenance and replacement strategies, requiring tracking the date of parts replacement by location for each asset across the portfolio. The resulting age calculation determines the replacement rate or percentage of parts replaced at any given operational age for that part type within that portfolio. Age function statistics can then be used to determine the optimal parts replacement strategy and to identify suboptimal maintenance processes within a site. In the case of large organizations with many different portfolios of assets, each with different parts, traditional manual processes for reliability engineering are only capable of analyzing a very small percentage of parts. Moreover, manually processing work orders to determine the age function is error-prone, and the sheer number of parts to monitor prevents the derived maintenance strategy from scaling or being approached holistically across the entire portfolio of operational assets.
[0008] Due to the diversity of work instruction quality and types within and across industries, a solution has not been developed to automatically reconstruct part history across an asset portfolio to continuously determine an age function for each single part at any given time.
[0009] One alternative to the manual reliability engineering processes described above is to extract maintenance insights from equipment sensor or telemetry data and optimize maintenance and replacement strategies using predictive maintenance techniques. One example is Uptake®, which uses machine learning models and asset sensor data to predict when the next failure may occur, thereby enabling corrective maintenance.
[0010] In current predictive maintenance systems, sensor data is analyzed to detect anomalies in sensor data trends that indicate asset failure. A few of these systems can separate defect states for individual parts and their failure patterns, providing predictions of how far a part's condition is from the point of failure as it degrades. Thus, predictive maintenance can enable organizations to plan and pre-allocate resources for corrective maintenance. However, the main drawback of these predictive maintenance solutions is that they heavily rely on assets or machinery with sufficient sensors to be effective, and access to that data is frequently sampled. Given the high cost of installing sensors, generally only a subset of parts and machinery within an asset fleet are instrumented. Therefore, these conventional solutions cannot provide a comprehensive parts maintenance strategy across a deployed portfolio of assets.
[0011] A second problem with conventional predictive maintenance systems is that they cannot account for sudden or unexpected failures. Current predictive maintenance solutions rely on the gradual degradation of parts from the onset of failure to the point of complete malfunction. However, many parts fail suddenly without warning and are not suitable for current predictive maintenance analysis. Therefore, current predictive maintenance solutions can, at best, only address a subset of parts within an asset portfolio and cannot be used to adequately optimize and manage a complete portfolio of parts maintenance.
[0012] Therefore, it is necessary to be able to efficiently calculate age functions for each set of parts in an asset portfolio at scale in order to improve the quality of preservation process insights from transaction preservation records and to provide comprehensive preservation and replacement solutions for all parts in an asset portfolio. [Overview of the project]
[0013] At least the needs described above are addressed, and various embodiments of the present invention achieve technical solutions in the art.
[0014] According to a first embodiment of the present invention, a scalable automated maintenance optimization system for managing the replacement of parts in one or more assets of a portfolio comprises at least one computer-accessible storage device configured to store instructions, and at least one processor communicably connected to the at least one computer-accessible storage device and configured to execute instructions that, when executed, provide a plurality of program modules including a data acquisition module, a data integration module, and an age calculation module. The data acquisition module is configured to receive a plurality of work orders or notifications relating to one or more parts being replaced in one or more assets of a portfolio. The data integration module is configured to analyze the received plurality of work orders or notifications to generate an asset history, the asset history including information on the replacement of parts in one or more assets of a portfolio. The age calculation module is configured to calculate the replacement age for each part type in the portfolio.
[0015] In another embodiment of the system, the program modules further include a failure pattern analysis module configured to analyze the calculated replacement age for each part type in the portfolio in order to determine one or more failure patterns associated with each part type.
[0016] In another embodiment of the system, the program modules further include a parts replacement optimization module configured to determine a parts replacement strategy for each parts type based on at least one or more determined failure patterns associated with each parts type.
[0017] In another embodiment of the system, the program modules further include a user interface configured to provide the user with recommendations for managing parts replacement schedules and inventory, based at least on a determined parts replacement strategy.
[0018] In another embodiment of the system, the data acquisition module is further configured to extract at least one of the following from a plurality of received work instructions or notifications: item movement information, bill of materials information, sensor measurement readings, parts repair information, or parts cost information.
[0019] In another embodiment of the system, the data integration module is further configured to determine the location of the replaced part among the multiple parts of the same part type within an asset, when the asset includes multiple parts of the same part type and fewer parts than all of the multiple parts are replaced.
[0020] In another embodiment of the system, the data integration module is further configured to group functionally equivalent parts as the same part type when an asset includes multiple functionally equivalent parts having different identifiers.
[0021] In another embodiment of this system, the data integration module has parts that are functionally equivalent. When a part is replaced, the system is further configured to update the parts replacement schedule for the part being replaced. In another embodiment of the system, the data integration module is further configured to determine the functional importance of each part within the asset to the operation of the asset in the event of a part failure.
[0022] In another embodiment of the system, the age calculation module is further configured to calculate a separate replacement age for each part of the plurality of parts based on the position of each part within the asset when the asset includes a plurality of parts of the same part type.
[0023] In another embodiment of the system, the calculated age function includes a statistical distribution of the replacement ages of the plurality of parts of each part type within one or more assets of the portfolio. The statistical distribution of ages can be fitted to a Weibull function.
[0024] In another embodiment of the system, the failure pattern analysis module is further configured to determine a dominant failure pattern from one or more failure patterns associated with each part type based on the calculated age at that time. One or more failure patterns for each part type include one or more or one early degradation failure pattern, a random failure pattern, or a wear failure pattern.
[0025] In another embodiment of the system, the part replacement optimization module is further configured to perform one or more of an on-condition maintenance process or a part replacement strategy determination for each part type with a wear failure pattern.
[0026] In another embodiment of the system, the part replacement optimization module is further configured to determine a part replacement strategy for each part type based on the cost of replacement parts for each part type based on a Weibull output indicating that the failure pattern is a wear failure.
[0027] In another embodiment of the system, the part replacement optimization module is further configured to determine a part replacement strategy for each part type based on the predicted downtime of the asset due to failure of parts of the part type within the asset based on a Weibull output indicating that the failure pattern is a wear failure.
[0028] In another embodiment of the system, the parts replacement optimization module is further configured to perform one or more of root cause analysis, change of previous parts replacement strategies, or on-condition maintenance processes for each part type of the early failure pattern.
[0029] In another embodiment of the system, the parts replacement optimization module is further configured to perform an on-condition maintenance process for each part type of the random failure pattern.
[0030] In another embodiment of the system, a method for determining an inventory optimization strategy for replacement of parts within one or more assets of a portfolio, comprising receiving a plurality of work instructions or notifications related to one or more parts to be replaced within one or more assets of the portfolio, and analyzing the received plurality of work instructions or notifications to generate an asset history, the asset history including information related to replacement of parts within one or more assets of the portfolio, and calculating an age at replacement for each part type within the portfolio.
[0031] In another embodiment of the system, a non-temporary computer-readable storage medium is configured to store a computer-executed program for a scalable automated maintenance optimization system for managing the replacement of parts in one or more assets of a portfolio, the program comprising several program modules, including a data acquisition module, a data integration module, and an age calculation module. The data acquisition module is configured to receive several work orders or notifications relating to one or more parts being replaced in one or more assets of a portfolio. The data integration module is configured to analyze the received work orders or notifications to generate an asset history, which includes information about the replacement of parts in one or more assets of a portfolio. The age calculation module is configured to calculate the replacement age for each part type in the portfolio.
[0032] Furthermore, any or all of the methods and related features described herein may be implemented or performed by all or part of a device system, apparatus, or machine, including all or part of any of the systems, apparatus, or machines described herein, or a combination or partial combination thereof. [Brief explanation of the drawing]
[0033] Please understand that the attached drawings are intended to illustrate various embodiments and may include elements that are not at a fixed scale. Note that similar reference numerals in different drawings refer to the same object.
[0034] [Figure 1] This figure shows an example of a computing device system according to one embodiment of the present invention.
[0035] [Figure 2] This figure shows another example of a computing device system according to one embodiment of the present invention.
[0036] [Figure 3] This is a block diagram showing an asset aging calculation and predictive maintenance management system according to one embodiment of the present invention.
[0037] [Figure 4] This flowchart illustrates a method for automatically calculating asset age and generating maintenance and replacement recommendations according to one embodiment of the present invention.
[0038] [Figure 5] This block diagram shows an asset aging calculation and failure analysis system according to one embodiment of the present invention.
[0039] [Figure 6] This is a flowchart illustrating a parts failure maintenance decision-making method for a maintenance management system according to one embodiment of the present invention.
[0040] [Figure 7] A block diagram showing an example of the optimal replacement strategy capability of a predictive maintenance management system according to one embodiment of the present invention.
[0041] [Figure 8] A block diagram showing an assurance eligibility module for a predictive maintenance management system according to one embodiment of the present invention.
[0042] [Figure 9] This block diagram shows the data integration-based optimal parts inventory management capability of one embodiment of the present invention.
[0043] [Figure 10] This is a block diagram showing the maintenance plan revision capability of a predictive maintenance management system according to one embodiment of the present invention.
[0044] [Figure 11] This is a block diagram illustrating an extension of an objective benchmarking system for a predictive maintenance management system according to one embodiment of the present invention.
[0045] [Figure 12] This is a block diagram of a parts risk score calculation system according to one embodiment of the present invention.
[0046] [Figure 13] This figure shows an example of an age function related to block 316 for a part, according to one embodiment of the present invention.
[0047] [Figure 14a] This figure shows an example of a Weibull probability distribution for part types related to block 317 according to one embodiment of the present invention. [Figure 14b] This figure shows an example of a Weibull cumulative distribution function for part types related to block 317 according to one embodiment of the present invention.
[0048] [Figure 15] This figure shows an example of a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0049] [Figure 16] This figure shows an example of a report generation screen for a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0050] [Figure 17] This figure shows an example of a maintenance strategy modification screen for a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0051] [Figure 18] This figure shows an example of a survey options screen for a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0052] [Figure 19]This figure shows an example of a cause-and-effect diagram used for root cause analysis in a predictive maintenance management system according to one embodiment of the present invention.
[0053] [Figure 20] This figure shows an example of a graph-based root cause analysis method used in a predictive maintenance management system according to one embodiment of the present invention.
[0054] [Figure 21] This figure shows another example of a maintenance strategy modification screen for a graphical user interface for a predictive maintenance management system, according to one embodiment of the present invention.
[0055] [Figure 22] This figure shows an example of a maintenance analysis screen for a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0056] [Figure 23] This figure shows an example of a task list management screen for a graphical user interface for a predictive maintenance management system according to one embodiment of the present invention.
[0057] [Figure 24] This figure shows an example of a global site view for a predictive maintenance management system according to one embodiment of the present invention.
[0058] [Figure 25] This figure shows an example of a client site view for a predictive maintenance management system according to one embodiment of the present invention.
[0059] [Figure 26] This figure shows an example of a client-server architecture for a predictive maintenance management system according to one embodiment of the present invention. [Modes for carrying out the invention]
[0060] The present invention provides a system and method for efficiently integrating maintenance management records, automatically calculating the age of each part across a large asset portfolio, and generating predictive maintenance schedules and inventory management strategies. It should be noted that the present invention is not limited to these or any other examples provided herein, which are merely illustrative.
[0061] In this regard, the description herein includes several specific details in order to provide a complete understanding of the various embodiments of the invention. However, those skilled in the art will understand that the invention can be practiced at a more general level without one or more of these details. In other cases, well-known structures are not shown or described in detail in order to avoid unnecessarily obscuring the description of the various embodiments of the invention.
[0062] Throughout this specification, references to “one embodiment,” “an embodiment,” “exemplary embodiment,” “illustrated embodiment,” and “specific embodiment” mean that the specific features, structures, or characteristics described in relation to an embodiment are included in at least one embodiment. Therefore, the appearance of phrases such as “in one embodiment,” “in an embodiment,” “exemplary embodiment,” “in this illustrated embodiment,” and “in this particular embodiment” in this specification does not necessarily all refer to one embodiment or the same embodiment. Furthermore, specific features, structures, or characteristics of different embodiments may be combined in any preferred manner to form one or more other embodiments.
[0063] Unless otherwise explicitly stated or required by context, the word “or” is used in a non-exclusive sense in this disclosure. Furthermore, unless otherwise explicitly stated or required by context, the word “set” means one or more. For example, the phrase “set of things” means one or more of the things.
[0064] In the following description, some embodiments of the present invention may be implemented, at least in part, by a data processing device system configured by a software program. Such a program may be implemented equivalently as a plurality of programs, and some or all of such software programs may be constructed equivalently in hardware.
[0065] Furthermore, the phrase “at least” is sometimes used, or may be used, in this specification simply to emphasize the possibility that other elements may exist in addition to those explicitly enumerated. However, unless otherwise explicitly stated (such as by the use of the term “only”) or required by the context, the non-use of the phrase “at least” in this specification nevertheless includes the possibility that other elements may exist in addition to those explicitly enumerated. For example, the phrase “at least based on A” includes A, as well as the possibility of one or more other additional elements other than A. Similarly, the phrase “based on A” includes A, as well as the possibility of one or more other additional elements other than A. However, the phrase “based only on A” includes only A. Similarly, the phrase “configured to do at least A” includes a configuration for implementing A, as well as the possibility of one or more other additional actions other than A. Similarly, the phrase “configured to do A” includes a configuration for implementing A, as well as the possibility of one or more other additional actions other than A. However, the phrase “configured to do only A” includes only A. It means a configuration for the purpose of doing so.
[0066] The words “device,” “machine,” “system,” and the phrase “device system” all include one or more physical devices or subdevices (e.g., several pieces of equipment) that interact to perform one or more functions, regardless of whether such devices or subdevices are located in the same housing or in different housings. However, according to various embodiments, in order to exclude embodiments in which each device, machine, system, or device system resides across different housings, it may be explicitly specified that the devices or machines or device systems reside entirely in the same housing. The word “device” may be equivalently referred to as “device system” in some embodiments.
[0067] The term “Program” in this disclosure should be interpreted as including one or more programs, including as a set of instructions or modules that can be executed by one or more components in a system, such as a controller system or a data processing device system, to cause the system to perform one or more operations. The set of instructions or modules may be stored in any type of memory device, such as memory device systems 130, 251, or both, as will be described later, as shown in Figures 1 and 2, respectively. Furthermore, this disclosure may describe, or can describe, how the instructions or modules of a program are configured to cause the execution of an action. In this context, the phrase “configured to” includes at least (a) instructions or modules that are currently in a form executable by one or more data processing devices to cause the execution of an action (e.g., the instructions or modules are in a compiled, unencrypted form ready for execution), and (b) instructions or modules that are not currently in a form executable by one or more data processing devices but can be converted into a form executable by one or more data processing devices to cause the execution of an action (e.g., the instructions or modules are encrypted in a non-executable form but can be converted into a form ready for execution through the execution of a decryption process). Such a description should be considered equivalent to describing how instructions or modules are structured to cause the execution of an action. The word “module” can be defined as a set of instructions. The words “program” and “module” can each be interpreted as containing multiple subprograms or multiple submodules. In this regard, a reference to a program or module can be considered to refer to multiple programs or multiple modules.
[0068] Furthermore, it should be understood that information or data may be affected, manipulated, or transformed into different forms as it moves through various devices or workflows. In this regard, unless otherwise expressly stated or required by context, references to information or data herein include modifications to such information or data. For example, “Data X” may be encrypted for transmission, and references to “Data X” shall include both its encrypted and unencrypted forms unless otherwise required or indicated by context. Furthermore, the phrase “graphical user interface” as used herein includes visual representations presented through a display device system, which may include computer-generated text, graphics, animations, or one or more combinations thereof, which may include one or more visual representations initially generated by an image capture device, at least in part.
[0069] Furthermore, exemplary methods are described herein with respect to Figures 4 and 6. The diagrams are described as including blocks related to computer executable instructions. Note that each instruction related to such a block does not have to be a separate instruction, but can be combined with other instructions to form a combined instruction set. The same instruction set may be related to two or more blocks. In this regard, the block arrangements shown in method diagrams 4 and 6 of this specification are not limited to the actual structure of any program or instruction set, or the required ordering of method tasks, but rather, in some embodiments, such method diagrams 4 and 6 merely show the tasks that the instructions are configured to perform when executed by a data processing device system in conjunction with interaction with one or more other devices or device systems.
[0070] Furthermore, the block diagrams shown in Figures 3, 5, and 7-12 illustrate various program modules implemented by computer executable instructions. It should be noted that each instruction associated with such a program module does not necessarily have to be a separate instruction, but can be combined with other instructions to form a combined instruction set. The same instruction set may be associated with two or more program modules. In this regard, the program modules in the block diagrams shown in Figures 3, 5, and 7-12 herein are not limited to the actual structure of any program or instruction set, or the required ordering of method tasks. Such block diagrams shown in Figures 3, 5, and 7-12, in some embodiments, merely illustrate tasks that the instructions are configured to perform when executed by a data processing device system in conjunction with interaction with one or more other devices or device systems.
[0071] Figure 1 schematically shows system 100 according to several embodiments. In some embodiments, system 100 may be a computing device 100 (as shown in Figure 2). In some embodiments, system 100 includes a data processing device system 110, an input / output device system 120, and a processor-accessible memory device system 130. The processor-accessible memory device system 130 and the input / output device system 120 are communicatively connected to the data processing device system 110.
[0072] The data processing device system 110 includes one or more data processing devices that, together with other devices such as one or more of the devices in system 100, implement or execute control programs relating to some of the various embodiments. Each of the phrases “data processing device,” “data processor,” “processor,” and “computer” includes any data processing device, such as a central processing unit (”CPU”), a desktop computer, a laptop computer, a mainframe computer, a tablet computer, a personal digital assistant, a mobile phone, and any other device configured to process, manage, or handle data, whether implemented with electrical, magnetic, optical, biological, or other components.
[0073] The memory device system 130 includes one or more processor-accessible memory devices configured to store information, including information required to execute a control program relating to some of the various embodiments. The memory device system 130 may be a distributed processor-accessible memory device system, including multiple processor-accessible memory devices that are communicably connected to the data processing device system 110 via multiple computers and / or devices. On the other hand, the memory device system 130 does not have to be a distributed processor-accessible memory system and may therefore include one or more processor-accessible memory devices located within a single data processing device.
[0074] Each of the phrases “processor-accessible memory” and “processor-accessible memory device” includes any processor-accessible data storage device, whether volatile or non-volatile, electronic, magnetic or optical, or otherwise, including but not limited to registers, floppy disks, hard disks, compact disks, DVDs, flash memory, ROM (read-only memory), and RAM (random-access memory). In some embodiments, each of the phrases “processor-accessible memory” and “processor-accessible memory device” includes a non-temporary computer-readable storage medium. In some embodiments, the memory device system 130 may be considered a non-temporary computer-readable storage medium system.
[0075] The phrase “communicatively connected” includes any type of connection, whether wired or wireless, between devices, data processors, or programs whose data can be communicated. Furthermore, the phrase “communicatively connected” includes connections between devices or programs within a single data processor, connections between devices or programs located within different data processors, and connections between devices not located within any data processor at all. In this regard, although the memory device system 130 is shown separately from the data processing device system 110 and the input / output device system 120, those skilled in the art will understand that the memory device system 130 may be located entirely or partially within the data processing device system 110 or the input / output device system 120. Furthermore, although the input / output device system 120 is shown separately from the data processing device system 110 and the memory device system 130, those skilled in the art will understand that such a system may be located entirely or partially within the data processing system 110 or the memory device system 130, depending on the contents of the input / output device system 120. Furthermore, the data processing device system 110, the input / output device system 120, and the memory device system 130 may be located entirely within the same device or housing, or they may be located separately but communicately connected between different devices or housings. When the data processing device system 110, the input / output device system 120, and the memory device system 130 are located within the same device, the system 100 in Figure 1 may, in some embodiments, be implemented by a single application-specific integrated circuit (ASIC).
[0076] The input / output device system 120 may include a mouse, keyboard, touchscreen, another computer, or any device or combination of devices into which desired selections, desired information, commands, or any other data are input to the data processing device system 110. The input / output device system 120 may include any suitable interface for receiving information, commands, or any data from other devices and systems as described in various embodiments.
[0077] The input / output device system 120 may also include an image generation device system, a display device system, a speaker device system, a processor-accessible memory device system, or any device or combination of devices from which information, instructions, or any other data are output from the data processing device system 110. In this regard, if the input / output device system 120 includes a processor-accessible memory device, such a memory device may or may not form part of the memory device system 130. The input / output device system 120 may include any suitable interface for outputting information, instructions, or data to other devices and systems as described in various embodiments. In this regard, the input / output device system may include any suitable interface for outputting information, instructions, or data to various other devices or systems as described in various embodiments. It may include.
[0078] Figure 2 shows an example of a computing device system 200 according to several embodiments. In some embodiments, the computing device system 200 may include a processor 250 corresponding to the data processing device system 110 of Figure 1. The memory 251, input / output (I / O) adapter 256, and non-temporary storage medium 257 may, according to some embodiments, correspond to the memory device system 130 of Figure 1. The user interface adapter 254, mouse 258, keyboard 259, display adapter 255, and display 260 may, according to some embodiments, correspond to the input / output device system 120 of Figure 1. The computing device 100 may also include a communication interface 252 that connects to a network 253 for communicating with other computing devices 100.
[0079] Various methods 400 and 600 can be implemented by associated computer executable instructions in several exemplary embodiments. In various exemplary embodiments, a memory device system (e.g., memory device system 130) is communicably connected to a data processing device system (e.g., data processing device system 110, which may be referred to herein as "e.g., 110") and stores a program executable by the data processing device system to cause the data processing device system to perform various embodiments of methods 400 and 600. In these various embodiments, the program may include instructions that perform or cause to perform various instructions among the instructions relating to the execution of various embodiments of methods 400 and 600. In some embodiments, methods 400 and 600 may include associated blocks, or subsets of additional blocks other than those shown in Figures 4 and 6, respectively. In some embodiments, methods 400 and 600 may include different sequences shown between various associated blocks shown in Figures 4 and 6, respectively.
[0080] In some embodiments of the present invention, transactional maintenance records may be represented in work orders from an enterprise resource planning (ERP) system and stored in a maintenance record storage system. However, it should be noted that the type of maintenance record is not limited to the type of ERP system or a specific asset type or operating industry or geographical location.
[0081] Figure 3 shows a predictive maintenance management system 300, including an asset aging calculation system and a maintenance strategy optimization system 329, according to one embodiment of the present invention. In some embodiments of the present invention, maintenance records 302 from a client-based maintenance management system or other maintenance record storage system 301 may be provided to the asset aging calculation system 329, also known as IronMan®, through a data acquisition process 304.
[0082] In some embodiments, the maintenance record 302 may include one or more of the following: work instructions, notices, task lists 904, repair records, item transfers 305, measurement readings 306, or descriptions in the bill of materials 307. The item transfers 305 indicate which asset parts were removed from the asset for replacement. The measurement readings 306 indicate the number of units the asset was operating in at the time of part replacement (e.g., hours, miles, kilowatt-hours, fuel burn, or any other measurement relating to the duty cycle that has a dominant influence on age). The bill of materials 307 provides a detailed list of the items used in the part replacement. However, it should be noted that the maintenance record 302 may include other information or data that can be used to calculate asset age.
[0083] In some embodiments, the preservation record data is ported to the asset using the age function 316. Data integration processes 308 may be used to perform the provisional calculations required to accurately determine the age of various parts across the folio. The age function will be described in detail in later sections of this specification. In some embodiments, these data integration functions may include one or more of the following: a part replacement event classification function (part replacement event classifier) 312, a part location detection function (part location detector) 309, a functionally important component (FSC) prediction function (functional importance predictor) 310, or a functionally equivalent item prediction function (functional equivalent predictor) 311.
[0084] In some embodiments of the present invention, the system automatically constructs a hierarchy of parts and assets within a portfolio, in the form of an ontology 308B, from the asset level down to subassemblies and then to functionally important components (FSCs). This ontology is used for part allocation and by semantic rules to associate the physical location of parts with activities such as monetary payments, preventive maintenance, and corrective maintenance. When the system onboards a new customer with new asset types, or an existing customer with new asset types, the system automates the construction of a physical dependency hierarchy of assets to the system, to assemblies, or subsystems, down to functionally important components. FSCs may include parts that suffer failure modes whose impact has safety, operational, or economic implications and which may also be subject to preventive maintenance, as identified from work instruction data. The system may automatically adjust the FSC components as new data is received, splitting the current FSC into lower-level components or aggregating it into higher-level components to separate failure modes or set the most appropriate level for assigning preventive maintenance. The hierarchy is modeled as an ontology to give a level of associated semantic rules that can help identify the location of parts that an asset may have, such as two or more equivalent parts that fit into different locations.
[0085] In some embodiments, the parts replacement event classification function 312 classifies parts based on whether the parts were replaced as part of a planned maintenance event or an unplanned maintenance event, including whether the replacement event was a preventive or corrective action. In some embodiments, the parts location detection function 309 determines the location of the replaced parts based on information contained in, for example, the bill of materials 307, item movement 305, or work instruction description. For example, the parts replacement event classifier 312 may identify an event in which two wheels were replaced on a six-wheeled truck, and the parts location detection function 309 may identify which two of the six wheels were replaced. In some embodiments, the parts location detector 309 determines the location of the replaced parts based on descriptions in notifications contained in repair information 502 and history replacement information.
[0086] In some embodiments, the functional importance prediction function 310 determines whether a replaced part is important (from the asset's perspective) based on whether the asset can continue to operate despite the failure of the part. For example, an oil leak may not prevent a vehicle from operating, but a steering arm failure may prevent a vehicle from operating. In some embodiments, the functionally equivalent prediction function 311 mediates work orders 302 where equivalent parts are sourced from different manufacturers and therefore have different codes. Note that equivalent parts from the same manufacturer may also have different codes. The functionally equivalent predictor 311 prevents functionally equivalent parts from being grouped separately in the parts grouping process (parts grouper) 313, which can lead to inaccurate asset part aging calculations and incorrect parts replacement schedules and strategies.
[0087] The data integration function connects and enriches the maintenance event information stored in the maintenance record 302 so that all exchange events for a given part type can be aggregated by the data aggregator 315, before the calculation of the distribution of asset part ages by the age calculation function 316. This process takes into account the complexity of the assets, and therefore the tracking in the separate maintenance record 302. The problem is not straightforward due to the number of interconnected parts. Furthermore, this task is complicated by the fact that the result of the age calculation function 316 only makes sense if it is calculated for a part type based on its position on the asset. When a single part is attached to the asset, determining the position of the part to be replaced can be straightforward. However, when there are many instances of a single part type attached to the asset, determining the positions is often difficult because the work instruction 302 may refer to the number of parts to be replaced but not to their positions.
[0088] After one or more of the data integration functions 309, 310, 311, and 312 are applied to the data, part replacement events are grouped in a part grouping process 313 to form a complete asset history 314 of the parts in the asset, which is then aggregated through a data aggregator 315 across the entire portfolio of assets. The age of each asset in the portfolio can be determined via a measured reading 306, but the age of each part that can be replaced in the asset must be inferred based on the date of the historical replacement, which is only accurate if the part location is precisely determined by the part location detection function 309. In some embodiments of the present invention, information on the current age of parts that have not been replaced may also be obtained to more accurately determine the status of the parts in the portfolio. In some embodiments, information on parts that have not been replaced is included in the asset history 314.
[0089] In some embodiments of the present invention, the age calculation function 316 calculates the distribution of replacement ages for each part by part type across a portfolio of working assets collected by the data aggregator 315. In some embodiments, the age calculation function 316 first generates an age profile for each part. Figure 13 shows an example of an age profile for a part. As the asset is used, the age of a part increases. When the asset is idle, the age of a part may not increase, but this is not always the case. When a part is replaced, the age of the part is reset to 0. In some embodiments, the age calculation function generates an age profile for each part of the asset in the portfolio. The age profile also shows the current age of the part.
[0090] In some embodiments of the present invention, the age calculation function 316 further generates a distribution of replacement ages for all parts of the same part type within an asset. In some embodiments of the present invention, the same part type may refer to a broad category of parts, such as a truck wheel. In some embodiments of the present invention, the same part type may refer to a narrower category of parts, such as the left front wheel of a truck. In some embodiments of the present invention, the same part type may refer to an even narrower category of parts, such as the left front wheel of a truck driven on an unpaved road. In some embodiments of the present invention, the system allows the “part type” to be defined broadly or narrowly to adapt to different use cases for similar parts or assets within a portfolio.
[0091] In some embodiments, a fitter such as a Weibull fitter 317 may be applied to the distribution of replacement ages for all parts of a defined part type, obtained from the age calculation function 316. Figures 14a and 14b show examples of fitting a Weibull probability distribution function or a Weibull cumulative distribution function to the distribution of replacement ages for all parts, respectively. The calibrated Weibull function from the fitter 317 can then be used to determine the failure patterns of the parts using the failure pattern analysis module 322. It will be apparent to those skilled in the art that order probability or statistical functions can be used to generate a fit to the distribution of replacement ages for all parts. Failure metrics can be calculated using a metric calculator 323 based on the failure patterns of the parts. Other functions can be used instead of the Weibull function. However, it should be noted that this can be used to generate a distribution of ages at the time of part type replacement 317.
[0092] In some embodiments of the present invention, Weibull fitting may result in one or more of three parameters, namely shape, scale, and location, which can be used to construct probability and cumulative density functions associated with failure patterns. In some cases, if multiple failure patterns are identified for the same part, IronMan® may construct multiple parameter sets, each having its own failure pattern. The shape parameter allows for the identification of three basic failure patterns: premature failure, random failure, and wear failure. In some embodiments, the system provides recommendations for maintenance interventions associated with the failure pattern. For example, in the case of premature failure, if a previously determined replacement or maintenance strategy is to be stopped, a root cause analysis (RCA) may be performed, and on-condition maintenance may be considered if it is practical to stop unexpected premature failures until the cause is resolved. In the case of random failure, an on-condition maintenance process is applied if it is practical and cost-effective. In the case of wear failure, both scheduled replacement and on-condition maintenance are applicable, but scheduled replacement may be more advantageous when the shape parameter strongly indicates a wear failure pattern.
[0093] In some embodiments, the failure metrics calculated by the metric calculator 323 may be used to determine the optimal replacement strategy for each part type using the strategy optimization process 324. In some embodiments, the task list rationalization process 325 may rationalize the task list 904 so that each part is stocked and replaced cost-effectively and optimally. Insights and recommendations from the strategy optimization process 324 and the task list rationalization process 325 may be integrated by the output integration and write function 326 into an update form 303 stored, for example, in the client's maintenance management system. In some embodiments, the insights and recommendations generated by the output integration function 326 may be presented to the client user 328 via a user interface.
[0094] In some embodiments, the calibrated Weibull function 317 may be used to generate a reliability, availability, and maintainability (RAM) simulation 319, which effectively provides an overall system availability and downtime simulation at the asset or fleet level. By adding a cost element to the RAM simulation 319 and the Weibull function 317, a lifecycle cost simulation 318 may predict future costs based on existing maintenance processes and propose alternative results through a what-if analysis. The outputs from the RAM simulation and analysis 319 and the lifecycle cost simulation 318 may be used to generate a future parts replacement strategy (forward-thinking plan) 321 using an equipment repair and replacement prediction process 320. In some embodiments, a strategy optimization process 324 may utilize the information generated by the equipment repair and replacement prediction process 320. The forward-thinking plan 321 may supplement or enhance the insights and recommendations generated by the output integration function 326.
[0095] The IronMan® system 329 is an exemplary asset part aging calculation and predictive maintenance management system that includes the implementation of processes and functions 304-326 shown in Figure 3.
[0096] Figure 4 shows a method 400 for integrating maintenance records and automating the calculation of an age function, according to several embodiments of the present invention. In some embodiments, in step 405, Maintenance record data 302 is acquired using the data acquisition process 304. In step 410, work instruction data related to item movement 305 or bill of materials 307 is extracted from the maintenance record data 302. In step 415, measurement data 306 is extracted from the maintenance record data 302. The extracted data is then analyzed using data integration functions. Analysis and data integration step 430 performs a data integration process 308 using various data integration functions 309, 310, 311, and 312, and groups parts using a parts grouping process 313 to generate asset history 314. The data is then aggregated across the portfolio in step 425 using a data aggregator 315. In step 430, the replacement age distribution for each part type across the portfolio is calculated using an age calculation function 316. In step 435, failure metrics are derived by performing failure pattern analysis module 322 and metric calculation 323 on the derived replacement age distribution. In step 440, the optimal maintenance strategy is calculated using the strategy optimization function 324 and the task list rationalization function 325, based on at least the derived failure metrics. In step 445, maintenance process recommendations are generated.
[0097] Figure 5 shows a block diagram of a data integration and failure analysis system 500 applied to a single work order 302 according to one embodiment of the present invention. In some embodiments, the age calculation function 316 is initialized using information from a client maintenance management system or other maintenance database system 301, including all historical maintenance or work order records 302. In this regard, system 500 provides an illustrative outline of the analysis performed by the IronMan® system 329 at the level of a single work order 302. Detection of replaced parts by part location detector 309 and classification of replacements as planned or unplanned by part replacement event classifier 312 may be derived using a natural language processing algorithm for notifications contained in repair information 502, which may be recorded as plain text by the field worker performing the repair. In some embodiments, functionally equivalent parts may be predicted (using a predictor 311) based on item movement 305 and part cost 501 (since the costs of functionally equivalent parts often tend to be similar).
[0098]
[0099]
[0100] In some embodiments, functionally important parts (using function 310) are determined from item movement 305 and bill of materials 307, which excludes consumables from, for example, the age calculation function 316 analysis. In some embodiments, the age 503 of a part replaced in work instruction 302 is calculated using measured readings 306. If the measured readings do not include the working time of the part, which is often manually entered, the part age 503 may be estimated based on the asset age and the date of the last replacement. In some embodiments, measured readings for one or more parts may be provided by a sensor, and these values are included in the analysis. In some embodiments, an aggregator 313 may aggregate all replacement ages 503 of functionally equivalent parts across a portfolio of assets to generate the respective age calculation function 316. In some embodiments of the present invention, a Weibull distribution 317 is then fitted to the age function 316 to determine failure patterns.
[0101] Early failures 506, random failures 504, and wear-out failures 505 may each undergo a root cause analysis 508 to determine the cause of the failure, and an on-condition analysis 509 to support the investigation conducted by the root cause analysis 508. These analyses are particularly important when early or random failures are not expected but are predictable. This helps determine whether failures 506 and random failures 504 have been correctly diagnosed. However, in some embodiments, if failures are found to be early or random, they may be excluded from future calculations by the exclusion filter 507, as they require a different analysis. Often, the maintenance process is ineffective for random failures 504 and early failures 506 that may be attributable to the manufacturer 609 that installed the parts or to the maintenance team. The results of the exclusion filter 507 help inform inventory requirements used by the task list rationalization process 325. In the case of wear failures 505, root cause analysis 508 may help recommend an optimal time-based replacement strategy using the strategy optimization process 324.
[0102] In some embodiments, wear failures 505 may also be used to simulate future maintenance and operating costs using a lifecycle cost simulation 318, which also influences the optimized replacement strategy process 324. The optimized replacement strategy process 324, together with the identification of randomly failing parts 507, determines optimal inventory requirements and generates insights and recommendations for the user. In this regard, the predictive maintenance management system 300 generates replacement strategies aimed at increasing asset availability for beneficial use, improving reliability, and reducing downtime while reducing ownership costs by optimizing parts maintenance and replacement.
[0103] Figure 6 shows Method 600 for a decision-making process guiding the analysis of part failures, according to several embodiments of the present invention. A flowchart of Method 600 shows how warranty claims, part failures, and maintenance actions may be coordinated according to several embodiments of the present invention. In step 601, a part failure is recorded. In step 602, it is determined whether the part is eligible for a warranty claim. If so, in step 620, the original manufacturer (OEM) is contacted to make a warranty claim. If the part is not eligible for a warranty (No in step 602), in step 603, an initial analysis is performed to determine the failure pattern. In some embodiments, the failure pattern in step 603 is determined by performing a failure pattern analysis process 322. If the failure pattern is determined to be random in step 603, it may not affect the planning of a replacement strategy. In some embodiments, in step 615, inventory is optimized to have a sufficient number of replacement parts based on historical failures determined by the task list rationalization process 325. If it is determined in step 603 that the part failure is due to wear, in step 604 a functionally critical parts process 310 is performed to determine whether the asset can still be operated. If the asset cannot be operated (No in step 604), replacement is scheduled immediately in step 606. If the asset can be operated (Yes in step 604), it is determined in step 605 whether a planned replacement has already been scheduled. If no replacement is planned (No in step 605), replacement is scheduled in step 606. Otherwise (Yes in step 605), the replacement clock for the part is reset in step 607 to prevent the part from being replaced prematurely as part of a scheduled maintenance plan.
[0104] In step 608, if replacement is scheduled, a root cause analysis process 508 may be performed. In some embodiments, the root cause analysis process 508 may be performed if the failure pattern was identified as early in step 603, or if the frequency and impact of the failure were determined to be significant in step 603. In step 609, it is determined whether the root cause analysis 508 leads to any conclusions that may affect the maintenance schedule or failure prediction. In step 610, the results obtained from the root cause analysis process 508 performed in step 608 are implemented. In some embodiments, in the case of a potential warranty claim, the OEM may be contacted in step 620.
[0105] Figure 7 shows a block diagram of a system for determining an optimal replacement strategy based on a generated replacement age function, according to several embodiments of the present invention. A predictive maintenance management system 300, such as the IronMan® system 329, extracts, integrates, and processes transaction data and sensor data to create a parts replacement history for each part on an asset. In some embodiments, transaction data may be provided by maintenance records 302, and sensor data may be provided by measurement readings 306. In some embodiments, maintenance records 302 may include information about what maintenance events have occurred or will occur, such as item movement 305 and repair information 502. Maintenance records 302 may also provide context about the maintenance event (for example, if an engine fails, a new engine part may be ordered along with a description of the defect, such as an oil leak from cylinder head number 5 or a bearing failure in connecting rod number 10). In some embodiments, the measured reading 306 may include the duty cycle of a part derived from the sensor data-equipped asset (for example, the fuel burn rate may be calculated using the pulse width from each fuel injector supplying fuel to the engine, or the time the engine is running may be calculated using an engine speed sensor against a reference clock in the electronic control unit on the asset). The maintenance record 302 and the measured reading 306 provide supplementary information that is integrated to obtain the age function 316 (maintenance events and context from the maintenance record 302, as well as duty cycle information from the reading 306).
[0106] In some embodiments of the present invention, the predictive maintenance management system 300 evaluates the current parts replacement and maintenance strategy against the generated parts replacement history to determine the optimal replacement strategy. Determining the age at part replacement using the age calculation function 316 is important for generating the optimal replacement strategy in the strategy optimization process 324. The failure pattern analysis module 322 in the metric calculator 323 utilizes the replacement age of parts across the asset portfolio to derive failure metrics, which are required to determine whether the part type is eligible for the planned maintenance strategy 704. If the part type is eligible for the planned maintenance strategy, the failure pattern analysis module 322 determines the potential replacement timing. Optimization of replacement timing in the strategy optimization process 324 may be based on recognition of maintenance and downtime costs 703. In some embodiments, maintenance and downtime costs 703 may be determined from site operational characteristics 701 and cost metrics 501, as well as maintenance downtime metrics 702, which are derived from the asset history 314.
[0107] Figure 15 shows an exemplary graphical user interface 1500 for displaying a determined optimal replacement strategy according to several embodiments of the present invention. In some embodiments of the present invention, a user can select and view one or more assets for which they are responsible. In some embodiments of the present invention, the initial screen of the user interface 1500 shown in Figure 15 displays a list of parts for a specific asset (e.g., a trial track). In some embodiments, the displayed list of parts includes one or more of the following: part number 1510, part description 1520, part cost 1530, current maintenance strategy 1540, recommended maintenance strategy 1550, and user action 1560.
[0108] Part number 1510 is a unique identifier for each part, which can be set by the user and available in the maintenance record 302, or set by the IronMan® system 300. Part description 1520 describes the part type and may include other information such as the OEM name and OEM part number. Part cost 1530 may include the cost of the part from a specific supplier, or the average cost of the part over a period of time or through multiple suppliers. Current maintenance strategy 1540 reflects the current parts replacement and maintenance strategy for each part type in the system. Recommended maintenance strategy 1 550 reflects the optimal parts replacement and maintenance strategy determined by the predictive maintenance management system 300 for each part type in the system. Users can use the graphical user interface 1500 to take several actions for each part type.
[0109] In some embodiments, the graphical user interface 1500 may use part hazard scores 1205, information from the Weibullfitter 317, and functionally important parts 310 to determine the order in which parts are displayed in the tables shown in the graphical user interface 1500. This feature assists the user in prioritizing tasks for actions within the graphical user interface 1500. For example, the system 300 may first identify all parts that have recommendations for changing the maintenance strategy. Then, from these identified parts, the system may further identify hazardous or functionally important parts and associate them with a higher priority for display in the top row of the table. Often, an asset may have thousands of parts, and using information such as part hazard scores 1205, information from the Weibullfitter 317, and functionally important parts 310 can greatly improve the usability of the graphical user interface. The user can filter, search, and sort the list of parts that have recommendations for changing the maintenance strategy. The user can also select a single part to view a more detailed list of part recommendations for that part (each part may have two or more recommendations), as shown in Figure 16. The data in the table can be exported for offline analysis. System 300 can also generate a load sheet that can be sent to the client worksite 1101 to update the maintenance strategy stored in the local or central maintenance transaction system. The report can be fed back to the transaction system (e.g., maintenance record storage system 301) to update its master data, including the task list 904.
[0110] Figure 16 shows an example of a report generation screen 1600 of the user interface 1500 according to several embodiments of the present invention. The report generation screen 1600 allows the user to generate a report showing how recommendations are derived from the part event history in order to provide technical justification and audit trails for downstream decisions.
[0111] In some embodiments, user action 1560 includes approval 1561, modification 1562, and investigation 1563. If the user agrees to the recommended parts replacement and maintenance strategy 1550, the user can accept it by clicking the approval button 1561. For example, as shown in Figure 15, the current maintenance strategy 1540 for the first part on the displayed list, part number 88891519, is to replace the part when it fails. The recommended maintenance strategy 1550 from the predictive maintenance management system 300 is to change this part at 13,000 hours. If the user finds this change acceptable, the user can click the approval button 1561 for this part, and the predictive maintenance management system 300 can automatically change the current maintenance strategy for this part from "operate until failure" to "replace at 13,000 hours". In some embodiments, selecting the approval button 1561 displays a report generation screen 1600 for the selected part.
[0112] Alternatively, the user can potentially change the system's parts replacement and maintenance recommendations by clicking the Modify button 1562 or the Investigate button 1563. Figure 17 shows an example of a maintenance strategy modification screen 1700 of the user interface 1500 according to several embodiments of the present invention. In some embodiments, the maintenance strategy modification screen 1700 includes a strategy modification interface 1710, information on replacement events 1720, and a distribution of replacement ages. This includes 317 and the strategy optimization result (cost function) 1730. In some embodiments, the strategy modification interface 1710 includes details describing the characteristics of the failure pattern and the justification for the recommended parts replacement and maintenance strategy 1550, information about the recommended parts replacement and maintenance strategy 1550, and various options for the user to modify the recommended parts replacement and maintenance strategy 1550. For example, if the recommended parts replacement and maintenance strategy 1550 is to replace parts at 20,000 hours based on the estimated downtime cost 1730, the user may modify the recommendation to replace parts at 30,000 hours instead of the recommended 20,000 hours, or choose a different replacement strategy such as operating until failure or on-condition maintenance.
[0113] In some embodiments, the graphical user interface allows the user to modify the exchange events 1720, which will be used for generating the distribution 317, through several mechanisms such as outlier detection or user-selectable date ranges. This interactive filtering of the exchange events 1720 timeline allows the user to perform what-if functions that can be used to identify value opportunities in root cause analysis. For example, in the exchange events 1720 shown in the example in Figure 17, the user may choose to exclude older events prior to April 2017 from the analysis. In response, the system 300 automatically reruns the age calculation function to generate a new theoretical Weibull cumulative distribution to exclude the older events and reruns the strategy optimization result 1730. Thus, the user can easily simulate the effects of various filtering criteria. Another what-if function may be provided by allowing the user to modify several inputs to the strategy optimization 1730, such as downtime costs 703 or the time it takes to complete work, maintenance, and downtime metrics 702. Using different inputs or factors changes the optimal strategy regarding the associated curves and costs. The user may base their decision to modify the recommended parts replacement and maintenance strategy 1550 on one or more of the following: replacement event information 1720, replacement age distribution 317, and strategy optimization results 1730. The replacement event information 1720 indicates when historical parts replacements were performed, the current strategy, the new strategy, and when predicted parts replacements will be performed under the new strategy. In some embodiments, the replacement event information 1720 is automatically updated to show the new strategy and predicted maintenance events when the user modifies the new strategy (recommended parts replacement and maintenance strategy 1550) using the strategy modification interface 1710. In some embodiments, the replacement age distribution 317 is fitted to a Weibull function that shows a confidence measure of the calculated distribution, indicated by a shaded space around the predicted function curve 317, in addition to the cumulative failure distribution of parts over the life of the part.The strategy optimization result 1730 shows the optimization results for the selected parts based on factors such as replacement parts cost and downtime.
[0114] There may be several factors that contribute to the user's decision to modify the recommended parts replacement and maintenance strategy 1550. In some embodiments, the recommended parts replacement and maintenance strategy 1550 may be based on the minimum total cost predicted by the strategy optimization result 1730. However, a wider confidence range (width of the shaded area around the curve 317) may correlate with the sparsity of the historical parts replacement data. Human factors such as the availability of maintenance personnel to perform the parts replacement, safety concerns, or legal regulations may also influence the user's decision to modify the recommended parts replacement and maintenance strategy 1550.
[0115] In some embodiments of the present invention, the user can test different optimization factors (cost and time) to find an appropriate maintenance strategy for a part based on their risk preferences. In some embodiments, after the user modifies and accepts the recommended parts replacement and maintenance strategy 1550, a report is generated for the selected part. Surface 1600 is displayed.
[0116] Returning to Figure 15, in some embodiments, the user may select the Investigation button 1563 to initiate a Root Cause Analysis (RCA) before approving or modifying the recommended maintenance strategy 1550. In some embodiments, selecting the Investigation button 1563 causes the graphical user interface to display the Investigation Options screen 1800, shown in Figure 18. In some embodiments, the Investigation Options screen 1800 includes buttons for initializing the RCA and for modifying the current or recommended strategy using the Maintenance Strategy Modification screen 1700. In some embodiments, the Investigation Options screen 1800 is displayed in the form of a pop-up window.
[0117] In some embodiments of the present invention, selecting the “Start RCA” button on the investigation options screen 1800 initiates a root cause analysis process 508 to determine if there are any unexpected or undesirable underlying issues affecting the determined failure patterns for a part (early failure 506, random failure 504, and wear failure 505). Figure 19 shows an example of a simple cause-and-effect diagram 1900 that may be used to perform an RCA. The cause-and-effect diagram 1900 includes various factors grouped into different categories that may be selected to identify the root cause of a failure. For example, in the example shown in Figure 19, the root cause analysis may identify design and manufacturing issues and environmental issues as potential causes or contributing factors related to the failure of the part. This process of identifying contributing factors may include breaking down the part to a further atomic level and understanding various design, use, and operation contexts. For example, an autonomous asset is less likely to be a fatigue contributing factor because it is not driven by people. The user can modify the cause-and-effect diagram 1900 to add new branches relating to new groupings of contributing factors and to add additional factors to each branch as various cause-and-effect chains are explored and navigated.
[0118] In some embodiments, the root cause analysis process 508 includes a graphic data structure for identifying and storing information about conditions and evidence. For example, as shown in Figure 20, a graph 2000 may be used to facilitate and record a more detailed RCA, including breaking down the cause into conditions and actions, enabling the association of evidence and insights with the RCA. Evidence may include rich media such as images, audio, or video. In the example shown in Figure 20, a traction motor is identified as having an early failure caused by a ground short circuit. Evidence of the cause is provided by an image of a burnt winding. The RCA process follows various characteristic factor chains to determine the underlying conditions that may contribute to the failure and potential corrective actions to eliminate these conditions or causes. In the example shown in Figure 20, the RCA process may query a database to determine the potential causes of the short circuit, i.e., motor overload or breakdown of electrical insulation. Further queries may be generated to obtain more detailed cause information, i.e., that the individual motor may be overloaded due to other motors in parallel being unavailable or the absence of a warning system. The RCA process can also identify corrective actions that can be taken to eliminate these causes, namely, the addition of an automatic shutdown control system, operator training, or warnings.
[0119] This information can be automatically generated using a root cause information database that stores the association between observed failures, the evidence supporting those failures, the conditions or causes contributing to those failures, and corrective actions. Transaction information, such as that stored in the client ERP system 301 and sensor information 306, may be used to query graphs and prune to generate estimated causes of observed failure patterns. In some embodiments, the root cause analysis process 508 may involve similar parts or similar failures. This may include determining whether an RCA has been performed previously. Transaction and sensor data related to the current failure of a part may be compared with historical information about previous RCAs to determine whether the current failure has a similar cause to a previous failure or whether a new RCA is needed to determine a potential new cause. The historical data may be filtered by a timeline to observe new urgent trends in changes in failure modes and percentages, when improvements may have been made to reduce undesirable causes of failure.
[0120] In some embodiments, the user may select the Strategy Modification button on the Investigation Options screen 1800 to display the Maintenance Strategy Modification screen 1700 shown in Figure 21. In the example shown in Figure 21, the system 300 determines a random failure pattern, and for that, the heuristic rule suggests that the correct action is to consider an on-condition maintenance strategy. However, the information on replacement events 1720 is sparse and includes two failure events that were very early in the life of the part. The thickness of the shaded confidence band of the distribution 317 also indicates that the sample size is small and the random failure modes predicted by the Weibull distribution may be incorrect. These observations may indicate that an RCA process should be performed to determine whether the early failure had other causes that may have been improved and require modification of the recommended maintenance strategy 1550. For example, if the early failure event was caused by operator error that could have been corrected by proper training, the new maintenance strategy 1550 might be to replace the part at 20,000 hours instead of the current 30,000 hours, as indicated by the calculated age of the most recent replacement date.
[0121] In some embodiments, the user may choose to generate a temporary, mitigating on-condition maintenance strategy to mitigate the impact of premature failure while RCA detection and prevention actions or remediation actions are being implemented. Once the root cause of premature failure is eliminated or reduced, a wear maintenance and parts replacement strategy may be employed.
[0122] In some embodiments, the graphical user interface 1500 includes various sorting and filtering criteria that the user can use to display the determined optimal replacement strategy for a part. The user may choose to export the displayed recommendations to a file, such as a .csv (comma-separated values) file, using the export button 1570. In some embodiments, the predictive maintenance management system 300 sorts the listed parts to match and prioritize the display of parts whose current maintenance strategy 1540 does not match the recommended maintenance strategy 1550. In some embodiments, the user interface 1500 may highlight parts whose current maintenance strategy 1540 does not match the recommended maintenance strategy 1550, use a different font, or display parts with change recommendations in a different part of the user interface than other parts whose recommended maintenance strategy has not changed. In some embodiments, the user can filter or sort the displayed list of parts based on their own criteria for prioritizing the risk level of the listed parts. The user may make their own judgments about what constitutes priority in a changing business context.
[0123] In some embodiments, the predictive maintenance management system 300 updates the maintenance plan 1003 by comparing the age of a functionally equivalent part with the age required by the part replacement schedule, in order to avoid wasting part life by unnecessarily changing a part when the part is to be replaced with a functionally equivalent part. Figure 22 shows an example of a maintenance analysis screen 2200 included in the graphical user interface 1500. In some embodiments, the maintenance analysis screen 2200 uses, for example, an age function 316 to match and prioritize the display of maintenance items 2210 that may be unnecessary and result in waste. The parts replacement schedule is sorted in descending order of the calculated percentage of wasted lifespan 2230, or in descending order of the scheduled replacement date 2220. In some embodiments, the user may choose an action 2240 to reset the age clock 1006 to skip a scheduled replacement or to investigate further.
[0124] In some embodiments, the predictive maintenance management system 300 determines a set of parts to be replaced from the asset history 314. The set of parts to be replaced is compared with at least one of the parts task list, parts inventory, or parts replacement schedule to determine the parts utilization rate in the asset portfolio. The task list rationalization process 325 adjusts the parts ordering schedule and task list based on the determined parts utilization rate to avoid over-ordering or under-ordering of parts. Figure 23 shows an example of a task list analysis screen 2300 included in the graphical user interface 1500. In some embodiments, the task list analysis screen 2300 sorts the parts utilization rate 2330 to match and prioritize the display of parts 2320 that may or may not be included in the task list, for example, using a series of filters 2310. In some embodiments, the user may select an action 2340 to modify the quantity of parts on the task list.
[0125] Figure 8 shows a block diagram for a warranty determination system 800 for determining warranty eligibility according to several embodiments of the present invention. Due to the lack of visibility in the client's maintenance management system 301 into when a part was last replaced and therefore what the age of the part is now, the information generated by the age calculation function 316 is important for determining whether a failed part is eligible for a warranty claim. The asset history 314 can provide useful information for making a warranty recommendation 803 by accurately determining the replacement age of parts that are flagged as failed 805 in the asset history 314. Warranty eligibility 810 may be determined by checking that a warranty recommendation 803 has not yet been claimed by comparing it with a claimed warranty 804. In some embodiments, warranty eligibility 810 is further based on warranty criteria 802 shown in the OEM database 801.
[0126] Figure 9 shows a block diagram for an optimal parts inventory determination system 900 according to several embodiments of the present invention. In some embodiments, the optimal parts inventory 903 is determined based on a task list rationalization process 325. The parts inventory 901 is compared with the current task lists 904 in the task list rationalization process 325 and the bills of materials 307 required to fulfill those task lists. The parts and content usage consistency analysis process 902 identifies which parts are overused or underused and therefore over-ordered or under-ordered, and may be in excess or under-stocked, based on the task lists 904 preserved by the task list rationalization process 325. If a part is not frequently needed or used and is in excess stock, the IronMan® system 329 may recommend, for example, that the part not be ordered very often or not at all. Conversely, if a part is regularly used and is under-stocked, the IronMan® system 329 may recommend that more of that part be ordered according to the task list rationalization process 325. The recommendations generated by the IronMan® system 329 can be used to determine the optimal parts inventory 903 for that site and to generate the optimal task list 905 even if the current task list 904 does not exist.
[0127] Figure 10 shows a block diagram of a system 1000 for updating maintenance plans according to several embodiments of the present invention. The revised maintenance plan 1003 assumes that a part failure 1010 occurred before the scheduled current maintenance plan 1002 and that the part is replaced. It is needed at any time. If the age of a part is not reset by the age clock reset process 1006, the part is very likely to be replaced prematurely under the current maintenance plan 1002. In some embodiments, the IronMan® system 329 supports this function and assists the user in prioritizing age clocks that should be reset by the age clock reset process 1006, and generates a revised maintenance plan 1003 based on an analysis of the current maintenance plan 1002 and part failures 1010 from the asset history 314. In some embodiments, the IronMan® system 329 includes detailed information in the parts database 1001 (through data integration 308) about how the parts are connected or functionally equivalent 311, so that part failures 1010 can be matched with the current maintenance plan 1002.
[0128] Figure 11 shows a system 1100 for objectively benchmarking multiple sites 1101 according to several embodiments of the present invention. Without objective performance and reliability metrics generated by the metric calculator 323, it is difficult for capital-intensive operations to benchmark those sites. Objective benchmarking 1103 is the first step towards continuous improvement and is important because it is difficult to implement at site 1101 due to quality issues with maintenance data 301 and fluctuating maintenance practices. In some embodiments, the IronMan® system 329 solves the quality issues of maintenance data 301 for multiple sites 1101 and provides the same performance and reliability metrics 1110 for each site 1101 to give objective benchmarking 1103 between sites 1101. Objective benchmarking 1103 allows sites 1101 to quickly identify where to concentrate efforts to improve performance and reliability metrics 1110, as well as to learn best practices from successful sites.
[0129] Figure 12 shows a system 1200 for calculating a part risk score 1205 according to several embodiments of the present invention. The system identifies functionally important components, while the part risk provides further information that can be used to prioritize corrective actions. Risk utilizes one or more of the following to calculate the risk score: the Weibull distribution 317, the frequency of work orders 1201, the part cost 501, the labor cost 1202, and the cost of lost production and disruption 1203. Several factors influence the risk score. The impact of a failure may be characterized by a mixture of the cost of the part, the labor cost for the remediation work, and the cost equivalent to the lost production time while the asset is unavailable. The frequency of failure events may be a combination of the B-20 age from the Weibull distribution and the number of preventive or remediation work orders per unit time. The detectability of a failure can be derived from asset history work orders and notifications, ranging in severity from "the failure was sudden and there was no prior warning to the operator or maintenance personnel" to "the failure exhibited many different failure symptoms as the condition slowly deteriorated over time, the symptoms were evident to operations or maintenance staff, and there was sufficient time before the final failure to plan, schedule, and commence corrective maintenance that would minimize disruption to operations." Examples of failure symptoms may include increased noise, heat, vibration, odor, or other visual indicators that suggest an impending failure.
[0130] In some embodiments, a part risk score 1205 may be generated using a risk score function 1204 derived from the part's work order frequency 1201, Weibull statistics generated using a Weibull fitter 317, the cost of the part 501, labor costs 1202 for servicing the work order, and the costs of lost production and disruption in case of failure 1203. The work order frequency 1201 is determined from processed work order data 302 and asset history 314. In some embodiments, the calculation of the risk score 1205 is performed using a Weibull fitter The B-20 information derived from 317 is used for weighting. The B-20 information relates to the age at which 20% of the part units in the population are likely to fail, which helps capture the adverse effects of early failure 506. Part costs 501 are extracted from the client ERP system 302. The part risk score 1205 is important because it adds a risk management layer to the optimal parts inventory preserved by the task list rationalization process 325 by considering the distribution, impact, and cost of failures in the inventory decision process.
[0131] Figure 24 shows an example of a global site view for a predictive maintenance management system according to one embodiment of the present invention. Global site operations for one or more clients may be connected to the IronMan® system 329 using, for example, a client-server architecture. In the example shown in Figure 24, there are four clients, each with multiple sites 1101. In some embodiments, sites (which may be mining areas, locomotive factories, etc.) are located in different geolocations globally and / or nationally. There may be many sites 1101 within a country. In some embodiments, data may be stored locally at each site or in a central client repository for each client. The arrows in Figure 24 represent data coming into the central repository. The IronMan® system 329 connects to a central repository 301 to perform maintenance analysis tasks. In some embodiments, there may be no central repository 301, and data is stored locally at each site. In these embodiments, the IronMan® system 329 connects to a local repository 301 to perform maintenance analysis tasks. In some embodiments, client data may be stored on an on-premises server or in the cloud, and the IronMan® system 329 works with either system. In some embodiments, client data remains isolated from other clients for security and intellectual property reasons, which may be specific client requirements. However, it is also easy to share data across clients, for example, to provide additional context or benchmarking analysis, which is a requirement from multiple clients.
[0132] Figure 25 shows an example of a client site view for a predictive maintenance management system according to one embodiment of the present invention. The exemplary site view in Figure 25 shows two exemplary assets 2510 (an excavator and a truck) in the upper left, which are operational and collecting sensor data while working. The sensor data is recorded and stored in the client repository 301. Asset 2510 (e.g., the truck in the lower left of Figure 25) can enter the maintenance area for maintenance work, with old part 2520 being removed and new part 2530 being moved. Transaction data icons indicate many things from when maintenance is triggered until the maintenance is actually performed (notifications, work orders, task lists). All of this information is then recorded in local client data 301 for that site. This client data 301 can then be supplied to the global client data repository 301 or directly accessed by the IronMan® system 329. As mentioned above, the IronMan® system 329 can also access multiple instances of local data when there is no global system (dotted line in Figure 25). Next, all of this data is fed into the IronMan® system for analysis, and then fed back to the client's offices and sites to improve maintenance.
[0133] Figure 26 shows an example of a client-server architecture for a predictive maintenance management system according to one embodiment of the present invention. More specifically, Figure 26 shows how, from the perspective of data communication, client data 301 is sent to the IronMan® system 329. An example of how data is transferred is shown. Client data 301 is transferred to the IronMan® system 329 via a secure connection, such as a secure file transfer via SFTP / HTTPS. The ETL process and data integration into the database are performed within the IronMan® system 329. In some embodiments, the IronMan® system 329 is deployed via a cloud service such as AWS and is accessible by the client via a web portal. Client access is secure and authenticated. The IronMan® analysis screen in the upper right of Figure 26 is the user interface for the client to interact with the IronMan® system 329.
[0134] Although the present invention is shown by the description of its embodiments, and embodiments are described in considerable detail, it is not the applicant's intention to limit the scope of the appended claims to such detail or to restrict them in any way. Further advantages and modifications will be readily apparent to those skilled in the art. Thus, the present invention in broader embodiments is not limited to the specific details illustrated and described, representative apparatus and methods, and exemplary examples. Accordingly, deviations from such details may be made without departing from the spirit or scope of the applicant's general inventive concept.
Claims
1. 1. A scalable automated maintenance optimization system for managing the replacement of multiple parts within one or more assets of a portfolio, comprising: at least one computer accessible storage device configured to store instructions; when communicatively connected to the at least one computer-accessible storage device and executed, a data ingestion module configured to receive a plurality of maintenance records related to the replacement of one or more parts of the plurality of parts in the one or more assets of the portfolio; a data integration module configured to analyze the received plurality of maintenance records to generate an asset history, the asset history including information regarding replacement of one or more parts of the plurality of parts within the one or more assets of the portfolio; and a data aggregator configured to aggregate the data across the portfolio; an age calculation module configured to calculate an age-at-replacement probability distribution for each part type of the plurality of parts in the one or more assets of the portfolio, the calculated age-at-replacement probability distribution comprising a statistical distribution of ages-at-replacement for a plurality of parts of each part type in the one or more assets of the portfolio; a part replacement optimization module configured to determine a part replacement strategy for each part type based at least on the calculated age-at-replacement probability distribution; a user interface control module configured to control a user interface to display a list of the plurality of parts and the respective determined parts replacement strategies, the user interface control module configured to automatically move any part of the plurality of parts whose determined parts replacement strategy differs from a current parts replacement strategy to a position closer to the top of the displayed list of parts in the user interface; At least one processor configured to execute instructions providing a plurality of program modules including: Preparation, The user interface control module includes: displaying a parts replacement strategy modification screen including the generated asset history for a part of the plurality of parts, the calculated age-at-replacement probability distribution for the part, and a cost function for the part, wherein the determined parts replacement strategy for the part corresponds to a minimum of the cost function; receiving user input to modify one or more maintenance events in the generated asset history; displaying an updated calculated age-at-replacement probability distribution and an updated cost function for the part; updating the determined part replacement strategy to correspond to a minimum of the updated cost function; and further configured to: A scalable automated maintenance optimization system.
2. the plurality of program modules further comprising a failure pattern analysis module configured to analyze the calculated age-at-replacement probability distributions for each part type in the portfolio to determine one or more failure patterns associated with each part type; the part replacement optimization module is configured to determine the part replacement strategy for each part type based at least on the determined one or more failure patterns associated with each part type.
2. The scalable automated maintenance optimization system of claim 1.
3. the data acquisition module is further configured to receive a plurality of sensor measurement readings related to the replacement of the one or more parts of the plurality of parts; the plurality of maintenance records including information regarding the replacement maintenance event of the one or more parts of the plurality of parts; the plurality of sensor measurement readings including information regarding a duty cycle associated with the replacement of the one or more parts of the plurality of parts; the data capture module is further configured to aggregate the plurality of maintenance records and the plurality of sensor measurement readings to generate the asset history.
2. The scalable automated maintenance optimization system of claim 1.
4. the data integration module is further configured to, when an asset includes a plurality of parts of a same part type and fewer than all of the plurality of parts are replaced, determine a location of a replaced part among the plurality of parts of the same part type within the asset.
2. The scalable automated maintenance optimization system of claim 1.
5. The data integration module is further configured to, when an asset includes multiple functionally equivalent parts having different identifiers, group the multiple functionally equivalent parts as a same part type; the data integration module is further configured to, when a part is replaced with a functionally equivalent part, update a part replacement schedule for the replaced part.
2. The scalable automated maintenance optimization system of claim 1.
6. the data integration module is further configured to determine a functional importance of each part within the asset to an operation of the asset upon failure of the part; The user interface control module is configured to: and automatically moving the parts of the plurality of parts that differ from the part replacement strategy of the first embodiment to a position closer to the top of the displayed list of parts in the user interface based at least on the functional importance of the parts.
2. The scalable automated maintenance optimization system of claim 1.
7. the age calculation module is further configured to, when an asset includes multiple parts of the same part type, calculate a separate age-at-replacement probability distribution for each part of the multiple parts based on a location of each part within the asset.
2. The scalable automated maintenance optimization system of claim 1.
8. The method of claim 7, further comprising: providing a fitter configured to use a probability or statistical function to generate a fit to a distribution of replacement ages for all parts of a defined part type.
2. The scalable automated maintenance optimization system of claim 1.
9. the statistical distribution of ages is fitted to a cumulative Weibull distribution; 2. The scalable automated maintenance optimization system of claim 1.
10. the failure pattern analysis module is further configured to determine a dominant failure pattern from the one or more failure patterns associated with each part type based on the calculated age at replacement; the one or more failure patterns for each part type include one or more premature degradation failure patterns, a random failure pattern, or a wear-out failure pattern; 3. The scalable automated maintenance optimization system of claim 2.
11. the part replacement optimization module is further configured to perform, for each part type of the wear-out failure pattern, one or more of an on-condition maintenance process or a determination of the part replacement strategy; 11. The scalable automated maintenance optimization system of claim 10.
12. The part replacement optimization module, when the determined failure pattern is a wear-out failure, Based on the cost of replacement parts for each part type, or based on a projected downtime of an asset due to failure of a part of the part type within the asset; and determining the part replacement strategy for each part type.
11. The scalable automated maintenance optimization system of claim 10.
13. the part replacement optimization module is further configured to perform, for each part type of the premature degradation failure pattern, one or more of a root cause analysis, a modification of a previous part replacement strategy, or an on-condition maintenance process.
11. The scalable automated maintenance optimization system of claim 10.
14. The data integration module: determining a set of parts to be replaced from the asset history; comparing the set of parts to be replaced to at least one of a parts task list, a parts inventory, or a parts replacement schedule to determine a parts utilization rate; adjusting a parts ordering schedule and the task list based on the determined parts utilization rate to avoid over-ordering or under-ordering parts; and further configured to:
2. The scalable automated maintenance optimization system of claim 1.
15. 1. A processor-implemented method for determining an inventory optimization strategy for replacement of a plurality of parts within one or more assets of a portfolio, the method comprising: receiving a plurality of maintenance records relating to the replacement of one or more parts of the plurality of parts in the one or more assets of the portfolio; analyzing the received plurality of maintenance records to generate an asset history, the asset history including information regarding replacement of one or more parts of the plurality of parts within the one or more assets of the portfolio; aggregating the data across the portfolios; and calculating a replacement age probability distribution for each part type of the plurality of parts in the one or more assets of the portfolio, the calculated replacement age probability distribution comprising a statistical distribution of replacement ages of a plurality of parts of each part type in the one or more assets of the portfolio; determining a part replacement strategy for each part type based at least on the calculated replacement age probability distributions; and controlling the user interface to display the list of the plurality of parts and the respective determined parts replacement strategies by automatically moving parts of the plurality of parts whose determined parts replacement strategy differs from a current parts replacement strategy to a position closer to the top of a displayed list of parts in the user interface; Including, Controlling the user interface includes: displaying a parts replacement strategy modification screen including the generated asset history for a part of the plurality of parts, the calculated age-at-replacement probability distribution for the part, and a cost function for the part, wherein the determined parts replacement strategy for the part corresponds to a minimum of the cost function; receiving user input to modify one or more maintenance events in the generated asset history; displaying an updated calculated age-at-replacement probability distribution and an updated cost function for the part; updating the determined part replacement strategy to correspond to a minimum of the updated cost function; Further comprising: Processor execution method.
