Prediction model enhancement

Abstract

An enhanced prediction model that utilizes product lifecycle segmentation and historical prediction data to generate more accurate future failure rates. Input data is segmented into groups based on the failure modes of the corresponding lifecycle. A prediction model such as Weibull analysis is implemented for each segmented group. The historical prediction data is also segmented into groups. The prediction parameters of each group of the segmented historical prediction data are compared to each other, and the comparison is then used to adjust the prediction parameters generated from the segmented groups of input data. Updated parameters for the input data are then output, thereby generating new future failure rates.

Description

Background Art

[0001] The present invention relates generally to failure rate prediction models, and more particularly to enhancing failure prediction models to make more accurate failure rate predictions.

[0002] Failure rate predictions are needed to provide assurance for systems and components. There are many types of predictive models for performing life data analysis. The failure rate of a system is typically time dependent, where the rate varies over the life cycle of the system or component. However, some predictive models are not well suited for some industries. In addition, the amount or quality of input data can affect the accuracy of failure predictions provided by the predictive model. For example, current methods for estimating component failure rates have limitations involving, for example, sample size, cold starts, and model fluctuations. In particular, the usefulness of failure estimates diminishes early in the life cycle of single component vintages, which skews future predictions. Therefore, these limitations reduce the usefulness of failure rate estimates.

[0003] In addition, it is often difficult to isolate the root cause of a product failure. Therefore, corrective action cannot be taken effectively. In addition, monitoring product performance history can provide reminders for things that have already happened, but it does not provide reminders for things that will happen in the future. For example, data used to predict a five-year return rate will only achieve high accuracy with a stable modeling output after collecting about two years' worth of data. Obtaining a stable and accurate prediction only after two years may be too late to subsequently request the supplier to take corrective action because the system or component may be at the end of its life. Every two or three years, the next generation of systems or components appears to have better capabilities. Thus, enhancing the accuracy of the failure rate predictions of the predictive model allows preventive actions to be taken more quickly to improve product reliability. Summary of the invention

[0004] It should be understood that this Summary is provided to introduce in simplified form a selection of concepts that are further described below in the Detailed Description. This Summary is not intended to be used to limit the scope of the claimed subject matter.

[0005] Embodiments of the present invention are directed to a computer-implemented method for enhancing a predictive model to generate future failure predictions. A non-limiting example of a computer-implemented method includes segmenting input data into groups. Implementing a predictive model for each group of the segmented input data. The computer-implemented method also includes segmenting historical prediction data into groups, and then comparing actual and predicted parameters within each group of the segmented historical prediction data. Based on the comparison, adjusting parameters defined by the predictive model for each group of the segmented input data. The computer-implemented method then includes outputting updated parameters of the predictive model for the input data.

[0006] According to another embodiment, a non-limiting example of a computer-implemented method includes inputting data into a predictive model to generate an initial failure mode for a first product. The initial failure mode provides a future failure rate of the first product defined by at least two parameters. The computer-implemented method also includes: segmenting historical prediction data, the historical prediction data including failure modes that have occurred during a life cycle of a second product; and then modifying the at least two parameters of the initial failure mode generated by the predictive model for the first product based on the segmented historical prediction data. Then, a new failure mode is generated using the predictive model, the predictive model using the modified at least two parameters to define a new future failure rate for the first product.

[0007] Embodiments of the present invention are directed to a system for enhancing a predictive model to generate future failure predictions. A non-limiting example of the system includes a processor communicatively coupled to a memory and a set of computer instructions stored in the memory, which, when executed by the processor, performs a method comprising: segmenting input data into multiple groups based on a life cycle, and each group of input data corresponds to a different failure mode of the life cycle. A predictive model is implemented for each group of segmented input data. Historical prediction data is segmented into groups based on failure modes that have occurred during the life cycle. The actual parameters and predicted parameters of each group of segmented historical prediction data are then compared. Based on these comparisons, the parameters defined by the predictive model are adjusted for each group of segmented input data. The method also includes outputting future failure rates using updated parameters of the predictive model. The method may then include storing the updated parameters as historical prediction data.

[0008] According to another embodiment, a non-limiting example of a system for enhancing a predictive model to generate future failure predictions includes a processor communicatively coupled to a memory and a set of computer instructions stored in the memory that, when executed by the processor, perform a method including inputting data into a predictive model to generate an initial failure mode for a first product. The failure mode provides a future failure rate of the first product defined by at least two parameters. The method then includes: segmenting historical prediction data, the historical prediction data including a plurality of failure modes that have occurred during a life cycle of a second product; and modifying the at least two parameters of the initial failure mode generated by the predictive model for the first product based on the segmented historical prediction data. Generate a new failure mode using the predictive model to define a new future failure rate for the first product using the modified at least two parameters.

[0009] Embodiments of the present invention are directed to a computer program product for generating future fault predictions, the computer program product comprising a computer-readable storage medium having program instructions embodied therewith. The program instructions are executable by a processor to cause the processor to perform a method. A non-limiting example of the method includes segmenting input data into groups based on a life cycle, and each group of segmented input data corresponds to a different fault. The method then includes implementing a prediction model for each group of segmented input data and segmenting historical prediction data into groups based on multiple failure modes. The actual parameters and predicted parameters of each group of segmented historical prediction data are compared with each other to determine differences that prove prediction errors. Based on these comparisons, the method includes adjusting parameters for each group of segmented input data defined by the prediction model. The method also includes outputting updated parameters of the prediction model. The method may also include storing the updated parameters as historical prediction data.

[0010] Additional technical features and benefits are achieved through the technology of the present invention. Embodiments and aspects of the present invention are described in detail herein, and these embodiments and aspects are considered to be part of the claimed subject matter. For a better understanding, reference is made to the specific implementation and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The details of the exclusive rights described herein are particularly pointed out and distinctly claimed in the claims at the end of the specification. The foregoing and other features and advantages of the embodiments of the present invention are apparent from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0012] Figure 1 depicts a block diagram of an exemplary computer processing system that may be used to implement exemplary embodiments of the present invention;

[0013] Figure 2 depicts a flow chart of an enhanced prediction model with different layers for generating future fault predictions according to one or more embodiments of the present invention;

[0014] Figure 3 depicts a generalized segmentation of input data based on lifecycle phases in accordance with one or more embodiments of the present invention;

[0015] Figure 4 Depicts a more specific segmentation of input data according to a particular lifecycle phase of a particular system, product, or component in accordance with one or more embodiments of the present invention;

[0016] Figure 5 According to a preferred embodiment of the present invention, Figure 2 A flowchart of the enhanced prediction model is provided;

[0017] Figure 6According to a preferred embodiment of the present invention, Figure 2 Enhanced Weibull prediction model for a specific system, product or component of the flowchart;

[0018] Figure 7 Describes how to calculate according to one or more embodiments of the present invention Figure 2 Flowchart of the learning layer of the equalization coefficients of the vector pairs;

[0019] Figure 8 Describes the use of the Weibull prediction model and calculation according to one or more embodiments of the present invention. Figure 2 The equalization coefficients of the vector pairs of the learning layer of the flowchart; and

[0020] Fig. 9 An exemplary process for enhancing a predictive model to provide more accurate failure predictions is depicted.

[0021] The figures described herein are illustrative. Without departing from the spirit of the present invention, there may be many variations of the figures or operations described herein. For example, these actions may be performed in different orders, or actions may be added, deleted or modified. Similarly, the term "coupling" and its variants describe having a communication path between two elements and do not imply a direct connection between these elements, without an intermediate element / connection between them. All these variations are considered as a part of the specification.

[0022] In the drawings and the following detailed description of the disclosed embodiments, the various elements shown in the drawings are provided with two or three digital reference numbers. In the case of minor exceptions, the leftmost digit of each reference number corresponds to the figure in which its element is first shown. DETAILED DESCRIPTION

[0023] Different embodiments of the present invention are described herein with reference to the relevant drawings. Without departing from the scope of the present invention, alternative embodiments of the present invention may be designed. Various connections and positional relationships (e.g., above, below, adjacent, etc.) are described between the elements in the following description and the accompanying drawings. Unless otherwise indicated, these connections and / or positional relationships may be direct or indirect, and the present invention is limited in this respect by the schematic diagram. Thus, the coupling of entities may refer to direct or indirect coupling, and the positional relationship between entities may be direct or indirect positional relationships. In addition, various tasks and process steps described herein may be incorporated into a more comprehensive program or process with additional steps or functions not described in detail herein.

[0024] The following definitions and abbreviations are used to interpret the claims and specification. As used herein, the terms "comprises," "comprising," "includes," "including," "has," "having," "contains," or "containing," or any other variation thereof, are intended to cover a non-exclusive inclusion. For example, a composition, mixture, process, method, article, or apparatus comprising a list of elements is not necessarily limited to only those elements, but may include other elements not expressly listed or inherent to such composition, mixture, process, method, article, or apparatus.

[0025] Additionally, the term “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment or design described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments or designs. The terms “at least one” and “one or more” may be understood to include any integer greater than or equal to one, i.e., one, two, three, four, etc. The term “plurality” may be understood to include any integer greater than or equal to 2, i.e., two, three, four, five, etc. The term “connected” may include both indirect “connections” and direct “connections.”

[0026] The terms "about", "substantially", "approximately" and variations thereof are intended to include the degree of error associated with the measurement of a particular quantity based on the equipment available at the time of filing this application. For example, "about" may include a range of ±8% or 5%, or 2% of a given value.

[0027] For the sake of brevity, conventional techniques related to making and using various aspects of the present invention may or may not be described in detail herein. In particular, various aspects of computing systems and specific computer programs for implementing the various technical features described herein are well known. Thus, for the sake of brevity, many conventional implementation details are only briefly mentioned herein, or omitted entirely, without providing well-known system and / or process details.

[0028] Turning now to an overview of the solution disclosed herein, one or more embodiments of the present invention address the above-mentioned shortcomings of the prior art by enhancing the prediction model in order to determine more accurate prediction parameters. The prediction model is enhanced by utilizing multivariate segmentation on new input data and then having the prediction model process each segmented group of input data separately. The prediction model is then further enhanced by segmenting the historical prediction data into groups and determining the equalization coefficient factor pairs for each segmented group of historical prediction data. The previous prediction parameters are compared with the corresponding actual parameters to determine where the error occurs over time. These equalization coefficient factor pairs are used as feedback to the equalization coefficient model, and the prediction parameters for the segmented input data determined by the prediction model are also fed back into the same equalization coefficient model. The equalization coefficient factor pairs are used to adjust or modify the prediction parameters of the segmented input data. Thus, modifying the prediction parameters of the segmented input data enhances the prediction model, resulting in improved failure rate accuracy.

[0029] The shortcomings of the prior art are preferably addressed by allowing the same fault prediction model to be used on a group of multiple partitions without the need to update the prediction model itself. The prediction parameters are updated or modified so as to regularize the prediction model over a lifetime across multiple partitions. In other words, regularization is applied at the partition level, thereby allowing the prediction model to remain stable across multiple partitions.

[0030] See also Figure 1 , shows an embodiment of a processing system 100 that is configured as a personal computer for implementing the teachings herein. In this embodiment, the system 100 has one or more central processing units (processors) 121a, 121b, 121c, etc. (collectively or generally referred to as processors 121). In one or more embodiments, each processor 121 may include a reduced instruction set computer (RISC) microprocessor. The processor 121 is coupled to a system memory 134 and various other components via a system bus 133. A read-only memory (ROM) 122 is coupled to the system bus 133 and may include a basic input / output system (BIOS) that controls certain basic functions of the system 100.

[0031] Figure 1An input / output (I / O) adapter 127 and a network adapter 126 coupled to the system bus 133 are further depicted. The I / O adapter 127 may be a small computer system interface (SCSI) adapter that communicates with the hard disk 123 and / or the tape storage drive 125 or any other similar component. The I / O adapter 127, the hard disk 123, and the tape storage device 125 are collectively referred to herein as a mass storage device 124. An operating system 140 for execution on the processing system 100 may be stored in the mass storage device 124. The network adapter 126 interconnects the bus 133 with an external network 136, enabling the data processing system 100 to communicate with other such systems. A screen (e.g., a display monitor) 135 is connected to the system bus 133 via a display adapter 132, which may include a graphics adapter and a video controller for improving the performance of graphics-intensive applications. In one embodiment, the adapters 127, 126, and 132 may be connected to one or more I / O buses connected to the system bus 133 via an intermediate bus bridge (not shown). Suitable I / O buses for connecting peripheral devices such as hard disk controllers, network adapters, and graphics adapters typically include public protocols such as Peripheral Component Interconnect (PCI). Additional input / output devices are shown connected to system bus 133 via user interface adapter 128 and display adapter 132. Keyboard 129, mouse 130, and speaker 131 are all interconnected to bus 133 via user interface adapter 128, which may include, for example, a super I / O chip that integrates multiple device adapters into a single integrated circuit.

[0032] In an exemplary embodiment, processing system 100 includes a graphics processing unit 141. Graphics processing unit 141 is a specialized electronic circuit designed to manipulate and modify memory to accelerate the creation of images in a frame buffer intended for output to a display. In general, graphics processing unit 141 is very efficient in manipulating computer graphics and image processing and has a highly parallel structure, which makes it more efficient than a general-purpose CPU in algorithms that complete the processing of large blocks of data in parallel.

[0033] Therefore, if Figure 1 As configured in FIG. 1 , system 100 includes processing capabilities in the form of processor 121, storage capabilities including system memory 134 and mass storage 124, input devices such as keyboard 129 and mouse 130, and output capabilities including speakers 131 and display 135. In one embodiment, system memory 134 and a portion of mass storage 124 together store coordinated Figure 1 The operating system of the functions of the different components shown in. Moreover, in one or more embodiments, the system may include or be connected to one or more databases or database management systems.

[0034] Figure 2 A flowchart 200 of an enhanced prediction model with different layers for generating more accurate future failure predictions is depicted. Routine layer 202 may be referred to as a prediction layer. Routine layer 202 includes dashed arrows that illustrate a conventional process for generating failure patterns for predicting failure rates. The conventional process extends directly from new data input 210 to prediction model 212.

[0035] In the traditional process, new data input skips the behavior segmentation 214 and goes directly to the prediction model 212. The prediction model 212 can be any known model that uses machine learning, pattern recognition, or statistical techniques to give a computer the ability to train and learn using data. Traditional outputs are provided directly from the prediction model 212. For example, the parameters output from the prediction model for estimating future failure rates are estimated.

[0036] However, the intelligent process according to an embodiment of the present invention includes new data input 210 through behavioral segmentation 214, so that the new data input 210 is segmented into a plurality of groups 1 to N. For example, the segmentation can utilize unique characteristics to classify the input data 210 into groups such as the geographic, demographic, psychographic, behavioral, or benefit being sought. Figure 3 In one or more of the embodiments shown, the input data may be segmented according to the stage of the life cycle. Figure 4 Specific embodiments are depicted that include segmenting input data 210 according to a stage of a specific life cycle of a specific system, product, or component. In other words, segmentation can utilize the stage within the life cycle of the system, product, or component to divide the input data 210 into different groups or different failure modes. In one or more embodiments, for example, a first product and a second product are different vintages, and the first product is newer than the second product. The term "vintage" refers to a grouping of individual units of a component, product, or system that were manufactured within the same time frame and have the same specifications (form / fit / function) and are expected to have similar performance characteristics.

[0037] Still see Figure 2 , the segmented groups 1 to N in the routine layer 202 are passed to the prediction model 212, which implements machine learning for each segmented group. Any prediction model can be used, especially those commonly used in a particular industry. The prediction model analysis produces a model coefficient vector that provides parameters for making failure rate predictions. In other words, Figure 2 In , the failure mode generated by the prediction model 212 can be referred to as the initial failure mode defined by the corresponding parameters of each set of new input data 210. Figure 5 In Figure 2 The generalized enhanced prediction model used in the flowchart 200 and in Figure 6 In Figure 2 The enhanced Weibull prediction model used in the flowchart 200 is shown.

[0038] Figure 5 and Figure 6 and Figure 2 The flowchart corresponds to that in Figure 5 and Figure 6 , depicts new input data 210 and behavior segmentation 214 along with the new input data segmented into four groups. Machine learning of the general prediction model 212 produces a model coefficient vector M for each segmented group. The model coefficient vector M depends on the specific prediction model 212.

[0039] In one or more embodiments, for example, when the set of input data is small, a Weibull prediction model may be used, such as Figure 6 As depicted in . Weibull analysis utilizes either two or three parameter analysis. Two parameter analysis produces parameters β, η which are often referred to as "Beta" and "Eta", respectively. Therefore, in the Weibull forecasting model, M is equal to the coefficient vector pair (Beta, Eta).

[0040] See again Figure 2 , passing the parameters of the prediction model 212 for each segmented group from the routine layer 202 to the feedback layer 206. The feedback layer 206 also utilizes Figure 2 The learning layer 204 of the flowchart 200 of FIG. The learning layer 204 begins with one or more databases storing historical prediction data 220. The historical prediction data 220 includes previously predicted parameters and corresponding actual parameters for failure modes previously predicted by the prediction model 212 of the routine layer 202. The parameters of the historical prediction data 220 define the history of failure rates associated with the product stored in the one or more databases.

[0041] The historical prediction data 220 is passed through the behavior segmentation 224, which segments the historical prediction data into groups 1 to N in a manner similar to the behavior segmentation 214 of the routine layer 202. The historical prediction data 220 can be grouped based on previous failure modes corresponding to a life cycle (such as the life cycle used by the behavior segmentation 214). For example, segmenting the input data into groups can include segmenting the input data into groups based on a life cycle, and each group of input data corresponds to a different failure mode of the life cycle. Segmenting the historical prediction data into groups can include segmenting the historical prediction data into groups based on a plurality of failure modes that have occurred during the life cycle.

[0042] Still see Figure 2In the learning layer 204 of the process 200, the historical prediction data 220 for each of the segmented groups 1 to N is then passed to the coefficient optimization model 228. The coefficient optimization model 228 for each segmented group 1 to N compares the previous prediction value derived from the historical prediction data 220 with the corresponding actual value derived from the historical prediction data 220 to generate the equalization coefficient factor pairs 1 to N, as shown in the layer 204. Figure 4 The general example of generally depicts an equalization coefficient matrix [A, B] for a prediction model, where A and B are used to adjust parameters provided by the prediction model 212 from the routine layer 202. Figure 6 Depicted is a coefficient optimization factor set (BA, EA, PA) for a three parameter Weibull prediction model for a particular component or product, such as a hard drive. PA, BA, and EA are coefficients used to equalize the Weibull model over different segments. The BA coefficient adjusts the traditional Weibull Beta coefficient. The EA coefficient adjusts the traditional Weibull Eta coefficient. The PA coefficient adjusts the model input, which in the case of electronic / computer components is the power-on time (POH) of the component. The equalization coefficient factor pairs 1 to N of the learning layer 204 are then provided to the equalization coefficient model 236 of the feedback layer 206. The prediction parameters from the prediction model 212 of the routine layer 202 are also provided to the equalization coefficient model 236 of the feedback layer 206.

[0043] like Figure 5 and Figure 6 As shown in , the difference or gap between the parameters defining the previous prediction and the actual parameters within each group of the segmented historical prediction data 210 is compared to determine whether there is an error in any of the parameters of the previous failure mode. The historical prediction data can be monitored and then the error in the predicted failure rate can be determined by comparing the parameters of the predicted failure rate with the parameters of the actual failure rate. The equalization coefficients defining the difference or gap are then used as feedback to generate updated parameters for the prediction model. In other words, the difference or gap is provided to the equalization coefficient model 236 for use in combination with the prediction parameters of the prediction model 212 of the routine layer 202 in order to minimize any possible errors in the prediction parameters of the prediction model 212. Thus, modifying the parameters of the failure mode generated by the prediction model of a particular product based on the segmented historical prediction data may include determining the difference between the parameters of the failure mode stored in the historical prediction data 220. More specifically, Figure 7 The general process of how to calculate the equalization coefficients of the vector pairs for the feedback layer 2 of flowchart 200 is depicted. Figure 8 The process of how to calculate the equalization coefficients of the vector pairs of feedback layer 2 when flowchart 200 utilizes the Weibull prediction model is then described. Therefore, the equalization coefficient model 236 provides an enhanced output so that the parameters of the prediction model 212 are modified or updated and thus become more accurate.

[0044] In one or more embodiments, outputting the updated parameters of the prediction model 212 includes generating future failure rates using the adjusted parameters. Likewise, in one or more embodiments, the output of the updated parameters from the equalization coefficient model 236 becomes feedback and is stored in the learning layer 204 as historical prediction data 220.

[0045] Likewise, the updated parameters of the prediction model 212 can be used to trigger actions in the manufacturing process, such as preemptive or corrective actions. For example, the updated parameters can be used to initiate instructions to an automated assembly system, line, or device. The assembly system, line, or device can be controlled or instructed to change, such as stop, speed up, or slow down. In another example, the updated parameters can result in instructions to remove a component or a batch of components from an assembly line by an automated assembly device. The updated parameters of the prediction model 212 can also be used in other types of processes.

[0046] Fig. 9 A flow chart of a method for enhancing a prediction model to more accurately generate future fault predictions according to one or more embodiments of the present invention is depicted. The method 900 includes a process block 902 for segmenting input data into groups. The process block 904 then includes implementing the prediction model 212 for each segmented group. The method 900 also includes a process block 906 for segmenting historical prediction data and a process block 908 for comparing actual parameters and predicted parameters within each segmented group at the coefficient optimization model 228. Next, the process 900 includes adjusting parameters of each group of segmented input data based on the comparison. The process 900 also includes outputting updated parameters of the prediction model for the input data.

[0047] Process 900 may also include performing a Weibull analysis on the prediction model. Process 900 may also include segmenting the input data into groups based on the life cycle, and each group of input data corresponds to a different failure mode of the life cycle. Process 900 may also include determining the difference between the actual parameters and the predicted parameters within each group of segmented historical predictions. Process 900 may also include using the difference between the actual parameters and the predicted parameters within each group of segmented historical predictions as feedback to generate updated parameters of the prediction model.

[0048] Additional processes may also be included. It should be understood that Fig. 9 The processes depicted in the present invention are illustrative and other processes may be added or existing processes may be removed, modified or rearranged without departing from the scope and spirit of the invention.

[0049] The present invention may be a system, method and / or computer program product of any possible integrated technical detail level. The computer program product may include a computer-readable storage medium (or medium) having computer-readable program instructions thereon, the computer-readable program instructions being used to cause a processor to perform aspects of the present invention.

[0050] A computer-readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of computer-readable storage media includes the following: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device (such as a punch card) or a raised structure in a groove with instructions recorded thereon), and any suitable combination of the above. As used herein, a computer-readable storage medium should not be interpreted as a transient signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated by a waveguide or other transmission medium (e.g., a light pulse by a fiber optic cable), or an electrical signal transmitted by a wire.

[0051] The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a corresponding computing / processing device, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, a wide area network, and / or a wireless network). The network may include copper transmission cables, optical transmission fibers, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in a computer-readable storage medium within the corresponding computing / processing device.

[0052] The computer-readable program instructions for performing the operation of the present invention can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, configuration data of integrated circuits, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and procedural programming languages, such as "C" programming language or similar programming languages. The computer-readable program instructions can be executed completely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer, partially on a remote computer, or completely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network (including a local area network (LAN) or a wide area network (WAN)), or can be connected to an external computer (for example, by using the Internet of an Internet service provider). In some embodiments, an electronic circuit (including, for example, a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA)) can execute a computer-readable program instruction to personalize the electronic circuit by utilizing the state information of the computer-readable program instructions, so as to perform aspects of the present invention.

[0053] Aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each box of the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0054] These computer-readable program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine that, when executed by a processor of the computer or other programmable data processing apparatus, creates a device for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions may also be stored in a computer-readable storage medium that can instruct a computer, a programmable data processing apparatus, and / or other device to function in a particular manner, such that a computer-readable storage medium having instructions stored therein includes an article of manufacture that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0055] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other apparatus that causes a series of operating steps to be performed on a computer, other programmable apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable apparatus, or other device implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0056] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to different embodiments of the present invention.To this end, each box in the flow chart or block diagram can represent a part of a module, segment or instruction, which includes one or more executable instructions for realizing the logical function of the specification.In some alternative embodiments, the function marked in the box may not occur in the order marked in the figure.For example, depending on the function involved, the two boxes shown in succession can actually be performed substantially at the same time, or these boxes can sometimes be performed in reverse order.It will also be noted that each box in the block diagram and / or flow chart and the combination of the boxes in the block diagram and / or flow chart can be realized by a system based on special-purpose hardware, and the system based on special-purpose hardware performs a specified function or action or performs a combination of special-purpose hardware and computer instructions.

[0057] The description of different embodiments of the present invention has been presented for the purpose of illustration, but is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terms used herein are selected to best explain the principles of the embodiments, practical applications, or technical improvements over technologies found in the marketplace, or to enable other persons of ordinary skill in the art to understand the embodiments described herein.

Claims

1. A computer-implemented method for enhancing a predictive model to generate future failure predictions, comprising: inputting data into a predictive model to generate an initial failure pattern for a first product, wherein the initial failure pattern provides a future failure rate for the first product defined by at least two parameters; segmenting historical prediction data, the historical prediction data including a plurality of failure modes that have occurred during a life cycle of the second product; modifying the at least two parameters of the initial failure mode generated by the predictive model for the first product based on the segmented historical predictive data; and A new failure mode is generated using the predictive model, the new failure mode defining a new future failure rate of the first product using the modified at least two parameters.

2. The computer-implemented method of claim 1 , further comprising: The input data is segmented into groups, and an initial failure pattern is generated for each group, wherein each initial failure pattern is defined by at least two parameters.

3. The computer-implemented method of claim 1 , wherein modifying the at least two parameters of the failure mode generated by the predictive model for the first product based on the segmented historical predictive data comprises determining a difference between parameters of at least one of the plurality of failure modes stored within the historical predictive data.

4. The computer-implemented method of claim 1 , wherein the predictive model performs one of the following: A two-parameter Weibull analysis; or A three-parameter Weibull analysis.

5. The computer-implemented method of claim 1, wherein the parameters of the historical prediction data define a history of failure rates associated with the second product stored in a database.

6. The computer-implemented method of claim 1, wherein the first product and the second product are of different vintages, and the first product is newer than the second product.

7. The computer-implemented method of claim 1, further comprising monitoring the historical prediction data and determining an error in the predicted failure rate associated with the second product by comparing parameters of the predicted failure rate to parameters of actual failure rates.

8. A system for enhancing a predictive model to generate future failure predictions, comprising: a processor communicatively coupled to the memory; as well as A set of computer instructions stored in the memory, when executed by the processor, the set of computer instructions performs a method, the method comprising: The input data is divided into a plurality of groups based on a life cycle, and each group of input data corresponds to a different failure mode of the life cycle; implementing a prediction model for each group of segmented input data; segmenting the historical prediction data into groups based on a plurality of failure modes that have occurred during the life cycle; Compare the actual parameters and predicted parameters of each group of segmented historical forecast data; Based on the comparison, adjusting the parameters defined by the predictive model for each set of segmented input data; and The updated parameters of the prediction model are used to output a future failure rate.

9. The system of claim 8, wherein: The method performed by the processor further includes storing the updated parameters as historical prediction data.

10. A system for enhancing a predictive model to generate future failure predictions, comprising: a processor communicatively coupled to the memory; as well as A set of computer instructions stored in the memory, when executed by the processor, the set of computer instructions performs a method, the method comprising: inputting data into a predictive model to generate an initial failure pattern for a first product, wherein the failure pattern provides a future failure rate for the first product defined by at least two parameters; segmenting historical prediction data, the historical prediction data including a plurality of failure modes that have occurred during a life cycle of the second product; modifying the at least two parameters of the initial failure mode generated by the predictive model for the first product based on the segmented historical predictive data; and A new failure mode is generated using the predictive model, the new failure mode defining a new future failure rate of the first product using the modified at least two parameters.

11. The system of claim 10, wherein modifying the at least two parameters of the failure mode generated by the predictive model for the first product based on the segmented historical predictive data comprises determining a difference between an actual parameter and a predicted parameter of at least one of the plurality of failure modes stored within the historical predictive data.

12. A system as described in claim 10, wherein the parameters of the historical prediction data define a history of failure rates associated with the second product stored in a database, and wherein the method performed by the processor also includes monitoring the historical prediction data and determining an error in the predicted failure rate associated with the second product by comparing the parameters of the predicted failure rate with the parameters of the actual failure rate.

13. A computer program product for enhancing a predictive model to generate future failure predictions, comprising: A computer readable storage medium having program instructions thereon for: The input data is divided into groups based on the life cycle, and each group of divided input data corresponds to different faults; Implementing a prediction model for each group of segmented input data; Segmenting historical prediction data into groups based on multiple failure modes; Comparing the actual parameters with the predicted parameters in each group of the segmented historical prediction data; Based on the comparison, adjusting the parameters defined by the predictive model for each set of segmented input data; and The updated parameters of the prediction model are output.

14. The computer program product of claim 13, wherein: The predictive model performs a Weibull analysis.

15. The computer program product of claim 13, wherein: Outputting updated parameters includes generating future failure rates using the adjusted parameters.

16. The computer program product of claim 13, wherein: Comparing the actual parameter and the predicted parameter for each group of the segmented historical prediction data includes determining a difference between the actual parameter and the predicted parameter corresponding to each group of the segmented historical prediction data.

17. The computer program product of claim 13, further comprising storing the updated parameters as historical prediction data.

Images