Fault Probability Evaluation Method, Device, Storage Medium and Server

Abstract

The present application discloses a fault probability evaluation method, device, storage medium and server. The method includes: obtaining feature data related to faults of an object to be evaluated in the current period and historical periods; determining a first fault probability of the object to be evaluated for a fault occurring after the current period based on the feature data in the current period and a first mapping relationship, wherein the first mapping relationship is used to represent the relationship between the feature data and the fault probability; determining a second fault probability of the object to be evaluated for a fault occurring after the current period based on the feature data in the historical periods, the first mapping relationship and a second mapping relationship, wherein the second mapping relationship is used to represent the relationship between the feature data and the fault probability at different time durations after the feature data appears; determining a third fault probability of the object to be evaluated for a fault occurring after the current period based on the first fault probability and the second fault probability. Thereby, not only can fault situations be discovered in advance, but also the accuracy is relatively high.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and particularly to a method, device, storage medium and server for evaluating failure probability. Background Art

[0002] An electric vehicle uses a battery as an energy source and relies on an electric motor to drive the wheels to rotate to achieve movement. In an electric vehicle, the safety of the battery plays a very important role.

[0003] In the related art, a mapping relationship between battery data within a fixed time window and the battery failure probability is constructed, and the failure probability of the battery to be detected is calculated based on the mapping relationship. The accuracy of this method is not high. Summary of the Invention

[0004] In view of the above problems, this application provides a method, device, storage medium and server for evaluating failure probability, which can not only detect failure situations in advance, but also has high accuracy.

[0005] In a first aspect, this application provides a method for evaluating failure probability, the method including: obtaining feature data related to failure of an object to be evaluated in the current period and historical periods; determining a first failure probability that the object to be evaluated will fail after the current period based on the feature data in the current period and a first mapping relationship, where the first mapping relationship is used to represent the relationship between the feature data and the failure probability; determining a second failure probability that the object to be evaluated will fail after the current period based on the feature data in the historical periods, the first mapping relationship and a second mapping relationship, where the second mapping relationship is used to represent the relationship between the feature data and the failure probability at different time durations after the feature data appears; determining a third failure probability that the object to be evaluated will fail after the current period based on the first failure probability and the second failure probability.

[0006] In the technical solution of the embodiments of this application, a first failure probability that the object to be evaluated will fail after the current period is determined based on the feature data in the current period and the first mapping relationship between the feature data and the failure probability, and a second failure probability that the object to be evaluated will fail after the current period is determined based on the feature data in the historical periods, the first mapping relationship and the second mapping relationship between the feature data and the failure probability at different time durations after the feature data appears, and a third failure probability, that is, the final failure probability, that the object to be evaluated will fail after the current period is determined based on the first failure probability and the second failure probability. In this way, not only can failure situations be detected in advance, but also the second mapping relationship is time-related, taking into account that as time goes by after the feature data appears, the influence on the failure probability decays, improving the accuracy of the failure probability.

[0007] In some embodiments, determining a second failure probability that the object to be evaluated fails after the current period based on the characteristic data of historical periods, the first mapping relationship, and the second mapping relationship includes: determining a fourth failure probability that the object to be evaluated fails after the historical period based on the characteristic data of the historical period and the first mapping relationship; determining a fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period based on the characteristic data of the historical period and the second mapping relationship; and determining the second failure probability that the object to be evaluated fails after the current period based on the fourth failure probability and the fifth failure probability. Thus, considering the influence of time, the characteristic data of different historical periods will correspond to different conditional probabilities, i.e., the fifth failure probability, for the current period, thereby improving the accuracy of the failure probability.

[0008] In some embodiments, determining a fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period based on the characteristic data of the historical period and the second mapping relationship includes: determining a sixth failure probability that the object to be evaluated fails within the preset duration after the historical period based on the characteristic data of the historical period and the second mapping relationship; and determining the fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period according to the sixth failure probability. Based on the probability of failure within the preset duration after the historical period, the probability of failure outside the preset duration can be determined.

[0009] In some embodiments, determining a third failure probability that the object to be evaluated fails after the current period based on the first failure probability and the second failure probability includes: determining a first probability that the object to be evaluated does not fail after the current period based on the first failure probability; determining a second probability that the object to be evaluated does not fail after the current period based on the second failure probability; and determining the third failure probability that the object to be evaluated fails after the current period according to the first probability and the second probability. By determining the probability of not failing, the probability of failing can be accurately determined.

[0010] In some embodiments, the historical period includes at least one consecutive historical period before the current period, and the second probability includes at least one; determining the third failure probability that the object to be evaluated fails after the current period according to the first probability and the second probability includes: determining the third failure probability that the object to be evaluated fails after the current period based on the first probability and at least one second probability. The historical period can include multiple ones. When there are multiple ones, the third failure probability that the object to be evaluated fails after the current period can be determined based on the second probabilities corresponding to the multiple historical periods, thereby further improving the accuracy of the failure probability.

[0011] In some embodiments, the method further includes: obtaining a first data volume of the feature data in the current period and a second data volume of the feature data in the historical period; determining a first probability correction amount of the first failure probability based on the first data volume and the first correction relationship, and determining a second probability correction amount of the second failure probability based on the second data volume, the first correction relationship, and the second correction relationship; wherein the first correction relationship is a relationship between the data volume of the feature data and the failure probability error determined based on the first mapping relationship, and the second correction relationship is a relationship between the data volume of the feature data and the failure probability error determined based on the second mapping relationship; correcting the first failure probability based on the first probability correction amount and correcting the second failure probability according to the second probability correction amount; determining a third failure probability that the object to be evaluated fails after the current period based on the corrected first failure probability and the corrected second failure probability. Since the data volumes of the feature data in different periods may be different, correcting the failure probability based on the data volume can further improve the accuracy of the failure probability.

[0012] In some embodiments, the historical period includes at least one consecutive historical period before the current period, and the method further includes: determining the failure probability distribution of the object to be evaluated according to the third failure probability that the object to be evaluated fails after each historical period in the at least one historical period; determining the historical time when the object to be evaluated fails based on the failure probability distribution. Thus, the failure probability distribution can be determined through the continuous failure probabilities on the time axis, so that the time when the failure actually occurred in history can be estimated.

[0013] In some embodiments, the feature data related to the failure includes at least one type of feature data, and the determining manner of the first mapping relationship includes: determining a third mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein the third mapping relationship is used to represent the relationship between the corresponding type of feature data and the failure probability; constructing the first mapping relationship based on the at least one third mapping relationship. Since the first mapping relationship is determined based on multiple types of feature data, compared with being determined based on a single type of feature data, the failure probability has a higher accuracy.

[0014] In some embodiments, the determining manner of the second mapping relationship includes: determining a fourth mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein the fourth mapping relationship is used to represent the relationship between the corresponding type of feature data and the failure probability at different time durations after the corresponding type of feature data appears; constructing the second mapping relationship based on the at least one third mapping relationship and the at least one fourth mapping relationship. Since the second mapping relationship not only considers the influence of multiple types of feature data on the failure, but also considers the influence of time factors on the failure, the accuracy of the failure probability is improved.

[0015] In a second aspect, the present application provides a computer-readable storage medium having a program stored thereon, and when the program is executed by a processor, the foregoing fault probability evaluation method is implemented.

[0016] In the technical solution of the embodiment of the present application, when the computer-readable storage medium executes the program, by implementing the foregoing fault probability evaluation method, not only can a fault situation be detected in advance, but also it is considered that after the feature data appears, as time goes by, the influence on the fault probability decays, improving the accuracy of the fault probability.

[0017] In a third aspect, the present application provides a server, including: a memory, a processor, and a program stored on the memory and executable on the processor, and when the processor executes the program, the foregoing fault probability evaluation method is implemented.

[0018] In the technical solution of the embodiment of the present application, when the server executes the program, by implementing the foregoing fault probability evaluation method, not only can a fault situation be detected in advance, but also it is considered that after the feature data appears, as time goes by, the influence on the fault probability decays, improving the accuracy of the fault probability.

[0019] In a fourth aspect, the present application provides a fault probability evaluation device, the device including: an acquisition module for acquiring feature data related to a fault of an object to be evaluated in a current period and a historical period; a first determination module for determining a first fault probability of the object to be evaluated having a fault after the current period based on the feature data in the current period and a first mapping relationship, where the first mapping relationship is used to represent the relationship between the feature data and the fault probability; a second determination module for determining a second fault probability of the object to be evaluated having a fault after the current period based on the feature data in the historical period, the first mapping relationship, and a second mapping relationship, where the second mapping relationship is used to represent the relationship between the feature data and the fault probability at different time durations after the feature data appears; and a third determination module for determining a third fault probability of the object to be evaluated having a fault after the current period based on the first fault probability and the second fault probability.

[0020] In the technical solution of the embodiment of the present application, based on the characteristic data of the current period and the first mapping relationship between the characteristic data and the failure probability, the first failure probability of the object to be evaluated for failure after the current period is determined, and based on the characteristic data of the historical period, the first mapping relationship, and the second mapping relationship between the characteristic data and the failure probability at different time durations after the appearance of the characteristic data, the second failure probability of the object to be evaluated for failure after the current period is determined, and based on the first failure probability and the second failure probability, the third failure probability of the object to be evaluated for failure after the current period, that is, the final failure probability, is determined. In this way, not only can the failure situation be discovered in advance, but also the second mapping relationship is time-related, taking into account that as time goes by after the appearance of the characteristic data, the influence on the failure probability decays, improving the accuracy of the failure probability.

[0021] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the specific implementation manners of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present application. And in all the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0023] Figure 1 is a schematic diagram of an application scenario of some embodiments of the present application;

[0024] Figure 2 is a schematic flowchart of a failure probability evaluation method of some embodiments of the present application;

[0025] Figure 3 is a schematic flowchart of obtaining the second failure probability of some embodiments of the present application;

[0026] Figure 4 is a schematic flowchart of obtaining the third failure probability of some embodiments of the present application;

[0027] Figure 5 is a schematic flowchart of correcting the third failure probability of some embodiments of the present application;

[0028] Figure 6 is a schematic flowchart of constructing the first mapping relationship of some embodiments of the present application;

[0029] Figure 7 is a schematic flowchart of constructing the second mapping relationship of some embodiments of the present application;

[0030] Figure 8 Schematic diagram of the process of the fault probability evaluation method according to some embodiments of the present application;

[0031] Figure 9 Schematic diagram of the structure of the fault probability evaluation device according to some embodiments of the present application. Detailed implementation manners

[0032] The embodiments of the technical solutions of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments are only used to illustrate the technical solutions of the present application more clearly, so they are only examples and cannot be used to limit the protection scope of the present application.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion.

[0034] In the description of the embodiments of this application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "a plurality" is more than two, unless otherwise specifically defined.

[0035] Referring to "embodiments" herein means that the specific features, structures or characteristics described in connection with the embodiments can be included in at least one embodiment of this application. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0036] In the description of the embodiments of this application, the term "and / or" is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0037] In the description of the embodiments of this application, the term "a plurality" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0038] In electric vehicles, the safety of the battery plays a very important role. In related technologies, a mapping relationship between battery data within a fixed time window and the battery failure probability is established, and the failure probability of the battery to be detected is calculated based on the mapping relationship. Specifically: First, the voltage, temperature, internal resistance, etc. of multiple storage batteries are obtained, and a time window and a sliding step size are set to obtain multiple time windows. Then, the test values, average values, variances, etc. of the voltage, temperature, internal resistance, etc. of the storage battery within each time window are calculated, and based on the test values, average values, variances, etc. of the voltage, temperature, internal resistance, etc. of the storage battery within multiple time windows, a mapping relationship between battery data and the battery failure probability is constructed, and then the failure probability of the battery to be detected is calculated based on the mapping relationship.

[0039] In the above method, although a mapping between features within a certain time window and the failure probability is performed, it does not consider that after a feature appears, its influence on the failure probability will decay over time, that is, the time factor is not taken into account, resulting in low accuracy of the failure probability calculation.

[0040] To solve the problem of low accuracy in calculating the failure probability, this application provides a failure probability evaluation method. By using the first mapping relationship between feature data and the failure probability, and the second mapping relationship between the feature data and the failure probability within different time periods after the feature data appears, the third failure probability, that is, the final failure probability, of the object to be evaluated after the current cycle is determined. This not only can detect the failure situation in advance, but also because the second mapping relationship takes into account the time factor, the obtained failure probability has high accuracy.

[0041] The failure probability evaluation method disclosed in the embodiments of this application can not only be used to determine the failure probability of the battery on the vehicle, but also can determine the failure probability of other components on the vehicle and other objects other than the vehicle.

[0042] For the convenience of description, the following embodiments take the failure probability evaluation method of an embodiment of this application for determining the failure probability of the battery on the vehicle as an example for illustration.

[0043] The failure probability evaluation method disclosed in the embodiments of this application can be applied to, for example Figure 1In the application environment shown. Among them, the vehicle 1000 communicates with the server 2000 through a network. The vehicle 1000 reports battery data and the battery failure time (this time is the time when the battery failure is detected) to the server 2000, and the server 2000 stores the battery data and the battery failure time in the database correspondingly. The server 2000 can obtain the battery data and the battery failure time from the database to determine the first mapping relationship between the characteristic data and the failure probability, and the second mapping relationship between the characteristic data and the failure probability within different time durations after the occurrence of the characteristic data. Furthermore, the server 2000 obtains the characteristic data of the battery at different times, and determines the failure probability of the battery based on the characteristic data at different times, the first mapping relationship, and the second mapping relationship. Among them, the vehicle 1000 includes but is not limited to pure electric vehicles, hybrid vehicles, range-extended vehicles, etc. The server 2000 can be implemented by an independent server or a server cluster composed of multiple servers.

[0044] According to some embodiments of the present application, referring to Figure 2 , the failure probability evaluation method includes:

[0045] S201, obtain the characteristic data related to failures of the object to be evaluated in the current period and historical periods.

[0046] Exemplarily, the object to be evaluated refers to the battery on the vehicle. The period refers to the time for evaluating the failure probability of the battery on the vehicle, which can be but is not limited to one hour, one day, or one week, etc. For example, when the period is one day, the failure probability of the battery on the vehicle can be evaluated once a day. The characteristic data related to failures refers to some data that can reflect the corresponding failures. For example, when the failure is the insulation failure of the battery, the characteristic data related to the failure includes but is not limited to the insulation value and current of the battery. Specifically, it can be the statistical values of the insulation value such as the average value, variance value, maximum value, and minimum value of the insulation value, and the statistical values of the current such as the average value, variance value, maximum value, and minimum value of the current.

[0047] Obtaining the characteristic data related to failures of the object to be evaluated in the current period and historical periods can be to obtain the statistical values of the insulation value, current statistical values, etc. of the battery within the day, and obtain the statistical values of the insulation value, current statistical values, etc. of the battery within the previous day, the previous two days,..., the previous n days, where n is a positive integer.

[0048] S203, based on the characteristic data in the current period and the first mapping relationship, determine the first failure probability of the object to be evaluated after the current period.

[0049] Among them, the first mapping relationship is used to represent the relationship between feature data and failure probability. This first mapping relationship can represent the relationship between a kind of feature data and failure probability, or can represent the relationship between multiple kinds of feature data and failure probability. For example, when the feature data related to the failure includes the insulation value statistic and the current statistic, this first mapping relationship can be the relationship with the failure probability under the joint action of the insulation value statistic and the current statistic.

[0050] After obtaining the feature data related to the failure of the object to be evaluated in the current period, the feature data related to the failure in the current period can be input into the first mapping relationship, and the probability of the object to be evaluated having a failure after the current period is calculated, denoted as the first failure probability. For example, when obtaining the insulation value statistic, current statistic, etc. of the battery on the same day, the insulation value statistic and the current statistic, etc. are input into the first mapping relationship, and the probability of the battery having a failure after the same day is calculated, denoted as the first failure probability.

[0051] S205. Based on the feature data of the historical period, the first mapping relationship and the second mapping relationship, determine the second failure probability of the object to be evaluated having a failure after the current period.

[0052] Among them, the second mapping relationship is used to represent the relationship between the failure probability and different time durations after the feature data appears. This second mapping relationship can represent the relationship between the failure probability and different time durations after a kind of feature data appears, or can represent the relationship between the failure probability and different time durations after multiple kinds of feature data appear. For example, when the feature data related to the failure includes the insulation value statistic and the current statistic, this second mapping relationship can be the relationship with the failure probability at different time durations after the insulation value statistic and the current statistic appear.

[0053] After obtaining the feature data related to the failure of the object to be evaluated in the historical period, based on the feature data related to the failure in the historical period, the first mapping relationship and the second mapping relationship, the probability of the object to be evaluated having a failure after the current period is calculated, denoted as the second failure probability. For example, when obtaining the insulation value statistic, current statistic, etc. of the battery within the previous day, within the previous two days,..., within the previous n days, for each day within the previous n days, the corresponding insulation value statistic and current statistic, etc. can be respectively input into the first mapping relationship and the second mapping relationship, two failure probabilities are calculated, and based on the two failure probabilities, the second failure probability of the battery having a failure after the current day can be calculated.

[0054] S207. Based on the first failure probability and the second failure probability, determine the third failure probability of the object to be evaluated having a failure after the current period.

[0055] After determining the first failure probability that the object to be evaluated will fail after the current cycle based on the feature data of the current cycle, and the second failure probability that the object to be evaluated will fail after the current cycle based on the feature data of the historical cycle, the failure probability that the object to be evaluated will fail after the current cycle can be calculated according to the first failure probability and the second failure probability, denoted as the third failure probability. For example, when determining the first failure probability based on the insulation value statistical value, current statistical value, etc. within the current day of the battery, and determining the second failure probability based on the feature data within the previous day, the previous two days,..., the previous n days of the battery, the probability that the battery will definitely not fail after the current day can be calculated according to the first failure probability and the second failure probability, and then the probability that the battery will fail after the current day can be calculated based on the probability of definitely not failing, denoted as the third failure probability, that is, the final failure probability.

[0056] In the above embodiments, through the first mapping relationship between the feature data and the failure probability, and the second mapping relationship between the feature data and the failure probability within different time durations after the appearance of the feature data, the third failure probability that the object to be evaluated will fail after the current cycle, that is, the final failure probability, is determined. In this way, not only can the failure situation be detected in advance, but also the second mapping relationship is related to time, considering the influence of the feature data on the failure probability over time, thereby effectively improving the accuracy of the finally determined failure probability.

[0057] According to some embodiments of the present application, with reference to Figure 3 , based on the feature data of the historical cycle, the first mapping relationship and the second mapping relationship, determining the second failure probability that the object to be evaluated will fail after the current cycle includes:

[0058] S301, based on the feature data of the historical cycle and the first mapping relationship, determine the fourth failure probability that the object to be evaluated will fail after the historical cycle.

[0059] After obtaining the feature data related to the failure of the object to be evaluated within the historical cycle, the feature data related to the failure within the historical cycle can be input into the first mapping relationship to calculate the probability that the object to be evaluated will fail after the historical cycle, denoted as the fourth failure probability. For example, when obtaining the insulation value statistical value, current statistical value, etc. within the previous day, the previous two days,..., the previous n days of the battery, for the previous day, the corresponding insulation value statistical value and current statistical value can be input into the first mapping relationship to calculate the probability that the battery will fail after the previous day, denoted as the fourth failure probability; similarly, the fourth failure probability that the battery will fail after the previous two days,..., the previous n days can be obtained.

[0060] S303. Based on the characteristic data of the historical period and the second mapping relationship, determine the fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period.

[0061] After obtaining the characteristic data related to failures of the object to be evaluated within the historical period, based on the characteristic data related to failures within the historical period and the second mapping relationship, the probability that the object to be evaluated fails outside the preset duration after the historical period can be calculated, denoted as the fifth failure probability. Among them, the preset duration can be the time interval between the current period and the historical period. For example, when obtaining the insulation value statistical values, current statistical values, etc. of the battery within the previous day, within the previous two days,..., within the previous n days, for the previous day, based on the insulation value statistical value and current statistical value of the previous day and the second mapping relationship, the probability that the battery fails outside one day after the previous day can be calculated, denoted as the fifth failure probability; for the previous two days, based on the insulation value statistical value and current statistical value of the previous two days and the second mapping relationship, the probability that the battery fails outside two days after the previous two days can be calculated, denoted as the fifth failure probability;...; and so on, the probability that the battery fails outside n days after the previous n days can be obtained, denoted as the fifth failure probability.

[0062] S305. Based on the fourth failure probability and the fifth failure probability, determine the second failure probability that the object to be evaluated fails after the current period.

[0063] After obtaining the fourth failure probability and the fifth failure probability, the second failure probability that the object to be evaluated fails after the current period can be calculated according to the fourth failure probability and the fifth failure probability. For example, for the previous day, after calculating the fourth failure probability that the battery fails after the previous day and the fifth failure probability that the battery fails outside one day after the previous day based on the insulation value statistical value, current statistical value, etc. of the battery within the previous day, the fourth failure probability and the fifth failure probability can be multiplied to obtain the second failure probability that the battery fails after that day; similarly, for each of the previous two days,..., the previous n days, a second failure probability that the battery fails after that day can be obtained.

[0064] In the above embodiments, the influence of time is considered. The characteristic data of different historical periods will correspond to different conditional probabilities, that is, the fifth failure probability, for the influence on the current period, thus improving the accuracy of the failure probability.

[0065] According to some embodiments of the present application, based on the characteristic data of the historical period and the second mapping relationship, determining a fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period includes: based on the characteristic data of the historical period and the second mapping relationship, determining a sixth failure probability that the object to be evaluated fails within the preset duration after the historical period; and according to the sixth failure probability, determining the fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period.

[0066] After obtaining the failure-related characteristic data of the object to be evaluated within the historical period, the failure-related characteristic data within the historical period can be input into the second mapping relationship to calculate the probability that the object to be evaluated fails within the preset duration after the historical period, denoted as the sixth failure probability, and based on the sixth failure probability, calculate the fifth failure probability that the object to be evaluated fails outside the preset duration after the historical period.

[0067] For example, when obtaining the insulation value statistical value, current statistical value, etc. of the battery within the previous day, within the previous two days,..., within the previous n days, for the previous day, the insulation value statistical value and current statistical value of the previous day can be input into the second mapping relationship to calculate the probability that the battery fails within one day after the previous day, denoted as the sixth failure probability, and then subtract the sixth failure probability from 1 to calculate the fifth failure probability; for the previous two days, the insulation value statistical value and current statistical value of the previous two days can be input into the second mapping relationship to calculate the probability that the battery fails within two days after the previous two days, denoted as the sixth failure probability, and then subtract the sixth failure probability from 1 to calculate the fifth failure probability;...; and so on, the fifth failure probability that the battery fails outside n days after the previous n days can be obtained.

[0068] In the above embodiments, based on the probability of failure within the preset duration after the historical period, the probability of failure outside the preset duration can be determined.

[0069] According to some embodiments of the present application, referring to Figure 4 , based on the first failure probability and the second failure probability, determining a third failure probability that the object to be evaluated fails after the current period includes:

[0070] S401, based on the first failure probability, determining a first probability that the object to be evaluated does not fail after the current period.

[0071] The first probability that the object to be evaluated does not fail after the current period can be calculated by subtracting the first failure probability from 1. For example, after obtaining the first failure probability that the battery fails after the current day, the probability that the battery does not fail after the current day can be obtained by subtracting the first failure probability from 1, denoted as the first probability.

[0072] S403. Determine a second probability that the object to be evaluated does not fail after the current cycle based on the second failure probability.

[0073] The second probability that the object to be evaluated does not fail after the current cycle can be calculated by subtracting the second failure probability from 1. For example, after obtaining the second failure probability that the battery corresponding to each day in the previous day, the previous two days,..., the previous n days fails after that day, the probability that the battery corresponding to each day in the previous day, the previous two days,..., the previous n days does not fail after that day can be obtained by subtracting the second failure probability from 1, which is denoted as the second probability.

[0074] S405. Determine a third failure probability that the object to be evaluated fails after the current cycle according to the first probability and the second probability.

[0075] Multiply the first probability and the second probability, and subtract the result of the multiplication from 1 to obtain the third failure probability that the object to be evaluated fails after the current cycle. For example, multiply the first probability corresponding to the current day and the second probability that the battery corresponding to each day in the previous day, the previous two days,..., the previous n days does not fail after that day, and then subtract the result of the multiplication from 1 to obtain the third failure probability that the battery fails after that day.

[0076] In the above embodiments, the probability of failure can be accurately determined by determining the probability of non - failure.

[0077] According to some embodiments of the present application, the historical cycle includes at least one consecutive historical cycle before the current cycle, and the second probability includes at least one; determining the third failure probability that the object to be evaluated fails after the current cycle according to the first probability and the second probability includes: determining the third failure probability that the object to be evaluated fails after the current cycle based on the first probability and at least one second probability.

[0078] It can be understood that the historical cycle can include one or more. When it is one, multiply the first probability by one second probability, and subtract the result of the multiplication from 1 to obtain the third failure probability that the object to be evaluated fails after the current cycle; when it is multiple, multiply the first probability by multiple second probabilities, and subtract the result of the multiplication from 1 to obtain the third failure probability that the object to be evaluated fails after the current cycle.

[0079] In the above embodiments, the historical cycle can include multiple. When it is multiple, the third failure probability that the object to be evaluated fails after the current cycle can be determined based on the second probabilities corresponding to multiple historical cycles, thereby further improving the accuracy of the failure probability.

[0080] According to some embodiments of the present application, with reference to Figure 5 , the method further includes:

[0081] S501, obtain the first data volume of the feature data in the current period and the second data volume of the feature data in the historical period.

[0082] It can be understood that the data volumes of the object to be evaluated in each period of the current period and the historical period may be different, and may also be different from the data volumes used when determining the first mapping relationship and the second mapping relationship. As a result, there may be a certain error in the determined third failure probability due to the difference in data volume. Based on this, when obtaining the feature data of the current period and the historical period, the data volume of the feature data is also recorded, and then the third failure probability is corrected based on the data volume, where the data volume can be in frames.

[0083] For example, a vehicle may travel for 1 hour in a certain day, and the data volume of the corresponding insulation value is 100 frames. Therefore, when obtaining the insulation value statistical value, it is calculated based on the data volume of 100 frames. And on another day, it travels for 10 hours, and the corresponding data volume is 1000 frames. Therefore, when obtaining the insulation value statistical value, it is calculated based on the data volume of 1000 frames. Obviously, the accuracies of the insulation value statistical values obtained with different numbers of frames are different, and thus the accuracies of the third failure probabilities determined based on the insulation value statistical values are also different. Therefore, the third failure probability can be corrected based on the data volume.

[0084] S503, based on the first data volume and the first correction relationship, determine the first probability correction amount of the first failure probability, and based on the second data volume, the first correction relationship, and the second correction relationship, determine the second probability correction amount of the second failure probability.

[0085] Among them, the first correction relationship is the relationship between the data volume of the feature data determined based on the first mapping relationship and the failure probability error, and the second correction relationship is the relationship between the data volume of the feature data determined based on the second mapping relationship and the failure probability error. For example, for the first correction relationship, after obtaining the first mapping relationship, the feature data with data volume A1 can be input into the first mapping relationship, so as to calculate the failure probability P corresponding to the data volume A1 A1 , and at the same time, an actual failure probability P can be obtained SA1 , and then obtain the difference △P between the two failure probabilities A1 = P A1 - P SA1 ; then input the feature data with data volume A2 into the first mapping relationship, so as to calculate the failure probability P corresponding to the data volume A2 A2 , and at the same time, an actual failure probability P can be obtained SA2 , and then obtain the difference △P between the two failure probabilities A2 = P A2 - PSA2 ;...; By analogy, a mapping table between the data volume and the difference in failure probability can be obtained. Furthermore, curve fitting can be performed based on the mapping table to obtain the first correction relationship. Similarly, the second correction relationship can be obtained.

[0086] After obtaining the first data volume of the characteristic data in the current period, the first data volume can be input into the first correction relationship to calculate the first probability correction amount of the first failure probability; after obtaining the second data volume of the characteristic data in the historical period, the second data volume can be input into the first correction relationship and the second correction relationship respectively to obtain the third probability correction amount and the fourth probability correction amount, and these two correction amounts can be used as the second probability correction amount of the second failure probability.

[0087] S505. Correct the first failure probability based on the first probability correction amount, and correct the second failure probability according to the second probability correction amount.

[0088] For example, subtract the first probability correction amount from the first failure probability to obtain the corrected first failure probability; subtract the third probability correction amount from the fourth failure probability to obtain the corrected fourth failure probability, and subtract the fourth probability correction amount from the sixth failure probability to obtain the corrected sixth failure probability. Furthermore, based on the corrected sixth failure probability, obtain the corrected fifth failure probability. For example, subtract the corrected sixth failure probability from 1 to obtain the corrected fifth failure probability, and based on the corrected fourth failure probability and the corrected fifth failure probability, obtain the corrected second failure probability. For example, multiply the corrected fourth failure probability and the corrected fifth failure probability to obtain the corrected second failure probability.

[0089] S507. Based on the corrected first failure probability and the corrected second failure probability, determine the third failure probability that the object to be evaluated will fail after the current period.

[0090] For example, subtract the corrected first failure probability from 1 to obtain the corrected first probability, and subtract the corrected second failure probability from 1 to obtain the corrected second probability. Then multiply the corrected first probability and the corrected second probability, and subtract the result of the multiplication from 1 to obtain the third failure probability.

[0091] In the above embodiments, since the data volumes of the characteristic data in different periods may be different, correcting the failure probability based on the data volume can further improve the accuracy of the failure probability.

[0092] According to some embodiments of the present application, the historical period includes at least one consecutive historical period before the current period, and the method further includes: determining the failure probability distribution of the object to be evaluated according to the third failure probability of the object to be evaluated after each historical period in at least one historical period; and determining the historical time of the object to be evaluated having a failure based on the failure probability distribution.

[0093] In the foregoing manner, the third failure probability corresponding to each historical period can be obtained. Based on the failure probability, the failure probability distribution of the object to be evaluated can be determined, such as a failure probability density curve. Then, based on the failure probability distribution, the historical time when the object to be evaluated has a failure can be determined, and this time is the time when the object to be evaluated actually has a failure.

[0094] For example, after obtaining the third failure probability corresponding to each historical period, the mathematical expectation of the historical time of having a failure can be calculated based on the third failure probability. Specifically, first, the third failure probabilities corresponding to each historical period are processed so that the sum of the processed failure probabilities is 1, and then the historical time of having a failure is calculated based on the formula:

[0095] E = t1*p1 + t2*p2 +... + t n *p n

[0096] Wherein, E represents the historical time of having a failure, t1 represents the first historical period, p1 represents the probability after processing the third failure probability corresponding to the first historical period, t2 represents the second historical period, p2 represents the probability after processing the third failure probability corresponding to the second historical period, and so on.

[0097] Exemplarily, t1 represents the previous day, represented by the number 1, t2 represents the day before the previous day, represented by the number 2,..., and so on, t n represents the nth day before, represented by the number n. Correspondingly, p1 is the probability after processing the third failure probability corresponding to the previous day, p2 is the probability after processing the third failure probability corresponding to the day before the previous day,..., and so on, p n is the probability after processing the third failure probability corresponding to the previous day. When processing the third failure probability, an equal-ratio processing method is adopted, and the sum of the processed probabilities is 1, that is, corresponding weights are assigned to the previous day, the day before the previous day,..., the nth day before based on the third failure probability. Then, based on the above formula, the historical time when the object to be evaluated actually has a failure can be calculated.

[0098] In the above embodiments, the failure probability distribution can be determined through the continuous failure probabilities on the time axis, so that the time when a failure actually occurred in history can be estimated.

[0099] According to some embodiments of the present application, the feature data related to the fault includes at least one type of feature data. Refer to Figure 6 , the determination method of the first mapping relationship includes:

[0100] S601, determine the third mapping relationship corresponding to each type of feature data in at least one type of feature data.

[0101] Wherein, the third mapping relationship is used to characterize the relationship between the corresponding type of feature data and the fault probability.

[0102] When there are multiple types of feature data, for each type of feature data, the relationship between the feature data and the fault probability can be determined, denoted as the third mapping relationship.

[0103] For example, when the feature data includes the insulation value statistic and the current statistic, for the insulation value statistic, a mapping table between the insulation value statistic and the fault probability can be established. Specifically, it can be a mapping table between the insulation value statistic in different insulation value statistic intervals and the probability of subsequent faults. Then, based on the mapping table, the third mapping relationship is constructed. Among them, since there is an inverse correlation between the insulation value statistic and the fault probability, multiple non-linear relationships such as -log(x), 1 / x, etc. can be obtained first, where the smaller the feature data, the higher the fault probability. Then, based on the mapping table between the insulation value statistic and the fault probability, multiple non-linear relationships are fitted to find the non-linear relationship with the smallest error as the third mapping relationship between the insulation value statistic and the fault probability.

[0104] In the same way, the third mapping relationship between the current statistic and the fault probability can be obtained. It should be noted that since there is a positive correlation between the current statistic and the fault probability, the obtained non-linear relationships are such as log(x), 1 - e(x), etc., non-linear relationships where the larger the feature data, the higher the fault probability.

[0105] S603, construct the first mapping relationship based on at least one third mapping relationship.

[0106] Based on multiple third mapping relationships, the first mapping relationship can be constructed by using the multiple regression method, that is, the relationship between the probabilities of subsequent faults under the action of multiple types of feature data is obtained.

[0107] In the above embodiments, since the first mapping relationship is determined based on multiple types of feature data, compared with being determined based on a single type of feature data, the fault probability has higher accuracy.

[0108] According to some embodiments of the present application, refer to Figure 7 , the determination method of the second mapping relationship includes:

[0109] S701, determine a fourth mapping relationship corresponding to each piece of feature data among at least one piece of feature data.

[0110] Among them, the fourth mapping relationship is used to characterize the relationship between the occurrence of the corresponding type of feature data and the failure probability within different time durations.

[0111] When there are multiple types of feature data, for each type of feature data, the relationship between the occurrence of this feature data and the failure probability within different time durations can be determined, denoted as the fourth mapping relationship.

[0112] For example, when the feature data includes insulation value statistical values and current statistical values, for the insulation value statistical values, a mapping table between the insulation value statistical values and the failure probability within different time durations after the occurrence of the insulation value statistical values can be established, and then the fourth mapping relationship is constructed based on the mapping table. Among them, although the probability of failure occurring every day after the occurrence of the feature data is decreasing, the sum of the probabilities of failure occurring over multiple days is increasing. Therefore, first obtain multiple non-linear relationships such as log(x), 1 - e(x), etc., where the longer the time, the greater the failure probability, but the increase in the failure probability becomes smaller. Then, fit multiple non-linear relationships to find the non-linear relationship with the smallest error as the fourth mapping relationship between the insulation value statistical values and the failure probability within different time durations after the occurrence of the insulation value statistical values. In the same way, the fourth mapping relationship between the current statistical values and the failure probability within different time durations after the occurrence of the current statistical values can be obtained.

[0113] S703, construct a second mapping relationship based on at least one third mapping relationship and at least one fourth mapping relationship.

[0114] For example, the at least one third mapping relationship and the at least one fourth mapping relationship can be multiplied to obtain the second mapping relationship.

[0115] In the above embodiments, since the second mapping relationship not only considers the influence of multiple types of feature data on failures, but also considers the influence of time factors on failures, the accuracy of the failure probability is improved.

[0116] As an example, take the insulation fault assessment of the battery on a vehicle as an example.

[0117] Refer to Figure 8 , the failure probability assessment method includes:

[0118] S801. Obtain the historical data of the battery from the database, and perform feature extraction on the historical data to obtain feature data related to the insulation fault of the battery, such as insulation values (insulation values under normal working conditions), current, temperature, insulation values when thermal management is turned on, and insulation values under other working conditions. Specifically, they can be insulation value statistical values, current statistical values, temperature statistical values, insulation value statistical values when thermal management is turned on, and insulation value statistical values under other working conditions.

[0119] S802. Construct the first mapping relationship between the feature data and the failure probability.

[0120] Taking the insulation value statistical value as an example. Obtain the probability of subsequent failures within different insulation value statistical value intervals of the insulation value statistical value, so as to establish a mapping table between the insulation value statistical value and the failure probability; obtain multiple non-linear relationships such as -log(x), 1 / x, etc., where the smaller the feature data, the higher the failure probability, and based on the mapping table between the insulation value statistical value and the failure probability, fit the multiple non-linear relationships to find the non-linear relationship with the smallest error, which is used as the third mapping relationship between the insulation value statistical value and the failure probability.

[0121] Using the same method, the third mapping relationship between the current statistical value and the failure probability, the third mapping relationship between the temperature statistical value and the failure probability, etc. can be obtained.

[0122] Then, based on multiple third mapping relationships, use the multiple regression method to construct the first mapping relationship, that is, obtain the relationship between the probabilities of subsequent failures under the action of multiple feature data.

[0123] S803. Construct the second mapping relationship between the feature data and the failure probability at different time intervals after the feature data appears.

[0124] Taking the insulation value statistical value as an example. Obtain the probability of failure within k days after the insulation value statistical value appears, such as the probability of failure within 1 day, the probability of failure within two days, etc., so as to obtain a mapping table between the insulation value statistical value and the failure probability at different days after the insulation value statistical value appears; obtain multiple non-linear relationships such as log(x), 1 - e(x), etc., where the longer the time, the greater the failure probability, but the increase in the failure probability becomes smaller, and then fit the multiple non-linear relationships to find the non-linear relationship with the smallest error, which is used as the fourth mapping relationship between the insulation value statistical value and the failure probability at different days after the insulation value statistical value appears.

[0125] Using the same method, the fourth mapping relationship between the current statistical value and the failure probability at different days after the current statistical value appears, the fourth mapping relationship between the temperature statistical value and the failure probability at different days after the temperature statistical value appears, etc. can be obtained.

[0126] Then, based on multiple third mapping relationships and multiple fourth mapping relationships, a fault condition probability mapping relationship, i.e., the second mapping relationship, is established as shown in the following formula:

[0127] P(Failure within k days|T) = W1 * [-log(x 1T )] + W2 * [-log(x 2T )] + … + W n *log(k) + b

[0128] where T represents the time when the characteristic data appears, P(Failure within k days|T) represents the probability that the battery has an insulation fault within k days after the characteristic data appears, W1, W2, …, W n are coefficients, x 1T , x 2T , etc. respectively represent the characteristic data 1, characteristic data 2, … corresponding to time T, -log(x 1T ), -log(x 2T ), etc. respectively represent the probabilities that the battery has an insulation fault within k days after the appearance of the characteristic data 1, characteristic data 2, etc., and b is a constant.

[0129] Furthermore, the probability of failure within k days P(Failure within k days) = P(Failure within k days|T) * P(T), and the probability of failure outside k days P(Failure outside k days) = 1 - P(Failure within k days).

[0130] S804. Obtain the real-time data of the battery from the database, and perform feature extraction on the real-time data to obtain the characteristic data related to the battery insulation fault. Taking one day as a cycle as an example, the characteristic data of the current day (t), the previous day (t - 1), the day before the previous day (t - 2), …, the nth day before (t - n) can be obtained. Among them, the characteristic data of each day can include the insulation value, current, temperature, insulation value when the thermal management is turned on, and insulation values under other working conditions. Specifically, it can include the statistical value of the insulation value, the statistical value of the current, the statistical value of the temperature, the statistical value of the insulation value when the thermal management is turned on, and the statistical value of the insulation value under other working conditions.

[0131] S805. Calculate the third fault probability that the battery fails after the current day.

[0132] Based on the characteristic data of the current day (t) and the first mapping relationship, calculate the first fault probability P(t) that the battery fails after the current day (t).

[0133] Based on the feature data of the previous day (t - 1) and the first mapping relationship, calculate the fourth failure probability P(t - 1) of the battery failing after the previous day (t - 1); at the same time, based on the feature data of the previous day (t - 1) and the second mapping relationship, calculate the sixth failure probability P(failure within 1 day|t - 1) of the battery failing within 1 day after the previous day (t - 1), as shown in the following formula:

[0134] P(failure within 1 day|t - 1) = W1 * [-log(x 1t-1 )] + W2 * [-log(x 2t-1 )] + … + W n *log(1) + b

[0135] Then, subtract the sixth failure probability P(failure within 1 day|t - 1) from 1 to obtain the fifth failure probability P(failure after 1 day|t - 1) of the battery failing after 1 day from the previous day (t - 1), that is, P(failure after 1 day|t - 1) = 1 - P(failure within 1 day|t - 1).

[0136] Next, multiply the fourth failure probability P(t - 1) and the fifth failure probability P(failure after 1 day|t - 1) to obtain the second failure probability of the battery failing after the current day (t), denoted as P(probability of failure after the current day, based on the feature data of the previous 1 day) = P(t - 1) * P(failure after 1 day|t - 1).

[0137] In the same way, based on the feature data of the previous two days (t - 2), the second failure probability of the battery failing after the current day (t) can be obtained, denoted as P(probability of failure after the current day, based on the feature data of the previous 2 days) = P(t - 2) * P(failure after 2 days|t - 2); and so on, based on the feature data of the previous n days (t - n), the second failure probability of the battery failing after the current day (t) can be obtained, denoted as P(probability of failure after the current day, based on the feature data of the previous n days) = P(t - n) * P(failure after n days|t - n).

[0138] Judging whether a failure will occur subsequently based on the feature data of different days is regarded as the superposition of the probabilities of different events on the current day. Therefore, the third failure probability of the battery failing after the current day (t) can be calculated by the following formula:

[0139] P(t z ) = 1 - [1 - P(t)] * Π[1 - P(probability of failure after the current day, based on the feature data of the previous i days)], where i ranges from 1 to n.

[0140] In the above manner, for the battery on each vehicle, a third failure probability can be obtained every day. Based on the third failure probability, insulation failure judgment can be performed on the battery on the vehicle, and the battery on the vehicle with a relatively large third failure probability can be regarded as the battery at risk of insulation failure, generating a corresponding list to provide a reference for taking market response measures to handle the failure.

[0141] S806. Determine the historical time when the battery fails based on the third failure probability.

[0142] For the battery on each vehicle, a third failure probability can be obtained every day. Based on the third failure probability, a failure probability density curve can be obtained, and then based on this curve, the historical time when the battery fails, that is, the expected failure time, can be determined.

[0143] In the above embodiments, based on the first mapping relationship between the feature data and the failure probability and the second mapping relationship between the feature data and the failure probability within different time periods after the appearance of the feature data, not only can the failure situation be detected in advance, but also the second mapping relationship is related to time, taking into account that after the appearance of the feature data, as time goes by, the influence on the failure probability decays, improving the accuracy of the failure probability. For example, the feature data in different cycles will multiply different conditional probabilities for the current cycle, which is equivalent to the farther the time, the smaller the influence. However, if the failure in a certain cycle is very serious, it is more likely to fail in the short term, and the corresponding decay is faster. At the same time, based on the continuous failure probabilities on the time axis, the failure probability distribution can be determined, such as calculating the expected failure time, so as to infer the time when the failure actually occurred in history.

[0144] According to some embodiments of the present application, a program is stored on a computer-readable storage medium, and when the program is executed by a processor, the foregoing failure probability evaluation method is implemented.

[0145] According to some embodiments of the present application, a server includes: a memory, a processor, and a program stored on the memory and executable on the processor. When the processor executes the program, the foregoing failure probability evaluation method is implemented.

[0146] According to some embodiments of the present application, with reference to Figure 9, the fault probability evaluation device 3000 includes: an acquisition module 3100, a first determination module 3200, a second determination module 3300, and a third determination module 3400. Among them, the acquisition module 3100 is used to acquire feature data related to faults of the object to be evaluated in the current cycle and historical cycles; the first determination module 3200 is used to determine a first fault probability that the object to be evaluated will have a fault after the current cycle based on the feature data in the current cycle and a first mapping relationship; wherein, the first mapping relationship is used to represent the relationship between the feature data and the fault probability; the second determination module 3300 is used to determine a second fault probability that the object to be evaluated will have a fault after the current cycle based on the feature data in the historical cycle, the first mapping relationship, and a second mapping relationship; wherein, the second mapping relationship is used to represent the relationship between the feature data and the fault probability at different time lengths after the feature data appears; the third determination module 3400 is used to determine a third fault probability that the object to be evaluated will have a fault after the current cycle based on the first fault probability and the second fault probability.

[0147] According to some embodiments of the present application, the second determination module 3300 is used to: determine a fourth fault probability that the object to be evaluated will have a fault after the historical cycle based on the feature data in the historical cycle and the first mapping relationship; determine a fifth fault probability that the object to be evaluated will have a fault outside the preset time length after the historical cycle based on the feature data in the historical cycle and the second mapping relationship; determine the second fault probability that the object to be evaluated will have a fault after the current cycle based on the fourth fault probability and the fifth fault probability.

[0148] According to some embodiments of the present application, the second determination module 3300 is used to: determine a sixth fault probability that the object to be evaluated will have a fault within the preset time length after the historical cycle based on the feature data in the historical cycle and the second mapping relationship; determine the fifth fault probability that the object to be evaluated will have a fault outside the preset time length after the historical cycle according to the sixth fault probability.

[0149] According to some embodiments of the present application, the third determination module 3400 is used to: determine a first probability that the object to be evaluated will not have a fault after the current cycle based on the first fault probability; determine a second probability that the object to be evaluated will not have a fault after the current cycle based on the second fault probability; determine the third fault probability that the object to be evaluated will have a fault after the current cycle according to the first probability and the second probability.

[0150] According to some embodiments of the present application, the historical cycle includes at least one consecutive historical cycle before the current cycle, and the second probability includes at least one; the third determination module 3400 is used to: determine the third fault probability that the object to be evaluated will have a fault after the current cycle based on the first probability and at least one second probability.

[0151] According to some embodiments of the present application, the apparatus 3000 further includes: a correction module (not shown) for obtaining a first data amount of the feature data in the current period and a second data amount of the feature data in the historical period; determining a first probability correction amount of the first failure probability based on the first data amount and the first correction relationship, and determining a second probability correction amount of the second failure probability based on the second data amount, the first correction relationship, and the second correction relationship; wherein the first correction relationship is a relationship between the data amount of the feature data and the failure probability error determined based on the first mapping relationship, and the second correction relationship is a relationship between the data amount of the feature data and the failure probability error determined based on the second mapping relationship; correcting the first failure probability based on the first probability correction amount and correcting the second failure probability according to the second probability correction amount; determining a third failure probability that the object to be evaluated fails after the current period based on the corrected first failure probability and the corrected second failure probability.

[0152] According to some embodiments of the present application, the historical period includes at least one consecutive historical period before the current period, and the apparatus 3000 further includes: a time estimation module (not shown) for determining the failure probability distribution of the object to be evaluated according to the third failure probability that the object to be evaluated fails after each historical period in at least one historical period; determining the historical time when the object to be evaluated fails based on the failure probability distribution.

[0153] According to some embodiments of the present application, the feature data related to the failure includes at least one type of feature data, and the first determination module 3200 is configured to: determine a third mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein the third mapping relationship is used to characterize the relationship between the corresponding type of feature data and the failure probability; constructing a first mapping relationship based on at least one third mapping relationship.

[0154] According to some embodiments of the present application, the second determination module 3300 is configured to: determine a fourth mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein the fourth mapping relationship is used to characterize the relationship between the corresponding type of feature data and the failure probability at different time durations after the corresponding type of feature data appears; constructing a second mapping relationship based on at least one third mapping relationship and at least one fourth mapping relationship.

[0155] For the relevant descriptions of the storage medium, the server, and the failure probability evaluation apparatus, please refer to the relevant descriptions of the failure probability evaluation method, which will not be elaborated here.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than limiting them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present application, and they should all be covered within the scope of the claims and the description of the present application. In particular, as long as there is no structural conflict, the various technical features mentioned in each embodiment can be combined in any way. The present application is not limited to the specific embodiments disclosed in the text, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for evaluating failure probability, characterized in that, The method includes: Obtaining feature data related to faults of an object to be evaluated in the current period and historical periods; Based on the feature data in the current period and a first mapping relationship, determining a first fault probability that the object to be evaluated will have a fault after the current period; wherein, the first mapping relationship is used to represent the relationship between feature data and fault probability; Based on the feature data in the historical periods, the first mapping relationship and a second mapping relationship, determining a second fault probability that the object to be evaluated will have a fault after the current period; wherein, the second mapping relationship is used to represent the relationship between the feature data and the fault probability at different time durations after the feature data appears; Based on the first fault probability and the second fault probability, determining a third fault probability that the object to be evaluated will have a fault after the current period; The method further includes: obtaining a first data volume of the feature data in the current period and a second data volume of the feature data in the historical periods; based on the first data volume and a first correction relationship, determining a first probability correction amount of the first fault probability, and based on the second data volume, the first correction relationship and a second correction relationship, determining a second probability correction amount of the second fault probability; wherein, the first correction relationship is the relationship between the data volume of the feature data determined based on the first mapping relationship and the fault probability error, and the second correction relationship is the relationship between the data volume of the feature data determined based on the second mapping relationship and the fault probability error; correcting the first fault probability based on the first probability correction amount, and correcting the second fault probability according to the second probability correction amount; based on the corrected first fault probability and the corrected second fault probability, determining a third fault probability that the object to be evaluated will have a fault after the current period.

2. The method according to claim 1, wherein The step of determining the second fault probability that the object to be evaluated will have a fault after the current period based on the feature data in the historical periods, the first mapping relationship and the second mapping relationship includes: Based on the feature data in the historical periods and the first mapping relationship, determining a fourth fault probability that the object to be evaluated will have a fault after the historical periods; Based on the feature data in the historical periods and the second mapping relationship, determining a fifth fault probability that the object to be evaluated will have a fault outside a preset time duration after the historical periods; Based on the fourth fault probability and the fifth fault probability, determining the second fault probability that the object to be evaluated will have a fault after the current period.

3. The method according to claim 2, characterized in that, The step of determining the fifth fault probability that the object to be evaluated will have a fault outside a preset time duration after the historical periods based on the feature data in the historical periods and the second mapping relationship includes: Based on the feature data in the historical periods and the second mapping relationship, determining a sixth fault probability that the object to be evaluated will have a fault within the preset time duration after the historical periods; According to the sixth fault probability, determining the fifth fault probability that the object to be evaluated will have a fault outside the preset time duration after the historical periods.

4. The method according to claim 1, wherein Determining a third failure probability that the object to be evaluated fails after the current period based on the first failure probability and the second failure probability includes: Determining a first probability that the object to be evaluated does not fail after the current period based on the first failure probability; Determining a second probability that the object to be evaluated does not fail after the current period based on the second failure probability; Determining a third failure probability that the object to be evaluated fails after the current period according to the first probability and the second probability.

5. The method according to claim 4, characterized in that, The historical period includes at least one consecutive historical period before the current period, and the second probability includes at least one; determining a third failure probability that the object to be evaluated fails after the current period according to the first probability and the second probability includes: Determining a third failure probability that the object to be evaluated fails after the current period based on the first probability and at least one of the second probabilities.

6. The method according to any one of claims 1-5, characterized in that, The historical period includes at least one consecutive historical period before the current period, and the method further includes: Determining the failure probability distribution of the object to be evaluated according to the third failure probability that the object to be evaluated fails after each historical period in the at least one historical period; Determining the historical time when the object to be evaluated fails based on the failure probability distribution.

7. The method according to any one of claims 1-5, characterized in that, The feature data related to the failure includes at least one type of feature data, and the determining method of the first mapping relationship includes: Determining a third mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein, the third mapping relationship is used to characterize the relationship between the corresponding type of feature data and the failure probability; Constructing the first mapping relationship based on at least one third mapping relationship.

8. The method according to claim 7, wherein The determining method of the second mapping relationship includes: Determining a fourth mapping relationship corresponding to each type of feature data in the at least one type of feature data; wherein, the fourth mapping relationship is used to characterize the relationship between the corresponding type of feature data and the failure probability at different time durations after the corresponding type of feature data appears; Constructing the second mapping relationship based on at least one third mapping relationship and at least one fourth mapping relationship.

9. A computer-readable storage medium, characterized in that, A program is stored thereon, and when the program is executed by a processor, it implements the failure probability evaluation method according to any one of claims 1-8.

10. A server, characterized in that, Including: A memory, a processor, and a program stored on the memory and executable on the processor, and when the processor executes the program, it implements the failure probability evaluation method according to any one of claims 1-8.

11. A fault probability evaluation device, characterized in that, The device includes: An acquisition module, configured to acquire feature data related to failure of the object to be evaluated in the current period and the historical period; A first determination module, configured to determine a first failure probability that the object to be evaluated fails after the current period based on the feature data of the current period and the first mapping relationship; wherein, the first mapping relationship is used to characterize the relationship between the feature data and the failure probability; A second determination module, configured to determine a second failure probability that the object to be evaluated fails after the current period based on the feature data of the historical period, the first mapping relationship, and the second mapping relationship; wherein, the second mapping relationship is used to characterize the relationship between the occurrence of feature data and the failure probability within different time durations; A third determination module, configured to determine a third failure probability that the object to be evaluated fails after the current period based on the first failure probability and the second failure probability; and Obtain a first data volume of the feature data of the current period and a second data volume of the feature data of the historical period; determine a first probability correction amount of the first failure probability based on the first data volume and the first correction relationship, and determine a second probability correction amount of the second failure probability based on the second data volume, the first correction relationship, and the second correction relationship; wherein, the first correction relationship is the relationship between the data volume of the feature data determined based on the first mapping relationship and the failure probability error, and the second correction relationship is the relationship between the data volume of the feature data determined based on the second mapping relationship and the failure probability error; correct the first failure probability based on the first probability correction amount, and correct the second failure probability according to the second probability correction amount; determine a third failure probability that the object to be evaluated fails after the current period based on the corrected first failure probability and the corrected second failure probability.

Images