CN110955226A

CN110955226A - Equipment failure prediction method and device, computer equipment and storage medium

Info

Publication number: CN110955226A
Application number: CN201911157656.3A
Authority: CN
Inventors: 杨明
Original assignee: Shenzhen General Interconnection Technology Co ltd
Current assignee: Shenzhen Huirong Iot Investment Enterprise LP
Priority date: 2019-11-22
Filing date: 2019-11-22
Publication date: 2020-04-03
Anticipated expiration: 2039-11-22
Also published as: CN110955226B

Abstract

The application relates to a device failure prediction method, a device, a computer device and a storage medium. The method comprises the following steps: acquiring a current operation data sequence corresponding to target equipment, wherein the current operation data sequence comprises a plurality of operation data corresponding to the target equipment in a current time period, and the operation data in the operation data sequence are sequenced according to operation time; acquiring a reference operation data sequence set, wherein the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal; respectively calculating the distance between the current running data sequence and each reference running data sequence in the reference running data sequence set to form a target distance set; and comparing the target distance set with the distance threshold value to obtain a comparison result, and obtaining a fault prediction result corresponding to the target equipment according to the comparison result. By adopting the method, the equipment failure prediction efficiency can be improved.

Description

Equipment failure prediction method and device, computer equipment and storage medium

Technical Field

The application relates to the technical field of internet of things, in particular to a method and a device for predicting equipment failure, computer equipment and a storage medium.

Background

With The development of computer technology, The Internet of Things (IOT) technology has emerged, which is a network that connects devices with The Internet according to an agreed protocol through information sensing devices such as radio frequency identification, infrared sensors, global positioning systems or laser scanners, etc., to exchange and communicate information, so as to realize intelligent identification, positioning, tracking, monitoring and management of devices. Because the equipment in the internet of things may have faults, people need to detect the faults of the equipment in the internet of things.

However, at present, manual fault detection is performed on the devices in the internet of things, and when more and more devices are in the internet of things, a large amount of manpower and time are consumed for fault detection of the devices in the internet of things, so that the device fault prediction efficiency is low.

Disclosure of Invention

In view of the above, it is necessary to provide a device failure prediction method, a device, a computer device, and a storage medium, which can improve the device failure prediction efficiency, in order to solve the problem of low device failure prediction efficiency in the above-described technology.

A method of device failure prediction, the method comprising: acquiring a current operation data sequence corresponding to target equipment, wherein the current operation data sequence comprises a plurality of operation data corresponding to the target equipment in a current time period, and the operation data in the operation data sequence are sequenced according to operation time; acquiring a reference operation data sequence set, wherein the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal; respectively calculating the distance between the current operating data sequence and each reference operating data sequence in the reference operating data sequence set to form a target distance set; and comparing the target distance set with a distance threshold value to obtain a comparison result, and obtaining a fault prediction result corresponding to the target equipment according to the comparison result.

In some embodiments, the comparing, according to the target distance set and the distance threshold, the reference operating data sequence set is an operating data sequence corresponding to an equipment fault, to obtain a comparison result, and obtaining a fault prediction result corresponding to the target equipment according to the comparison result, includes: calculating a minimum distance in the set of target distances; comparing the minimum distance with the distance threshold value to obtain a comparison result; and when the comparison result is that the minimum distance is smaller than the distance threshold, determining that the fault prediction result corresponding to the target equipment is a fault.

In some embodiments, the determining, by the processor, that the fault prediction result corresponding to the target device is a fault when the comparison result indicates that the minimum distance is smaller than the distance threshold includes: when the comparison result is that the minimum distance is smaller than the distance threshold, acquiring a reference operation data sequence corresponding to the minimum distance as a target reference operation data sequence; and acquiring a fault type corresponding to the target reference operation data sequence as a predicted fault type corresponding to the target equipment.

In some embodiments, the calculating the distance between the current running data sequence and each reference running data sequence in the reference running data sequence set respectively comprises: obtaining a distance matrix according to the current operation data sequence and the reference operation data sequence, wherein the current operation data sequence comprises n current operation data, the reference operation data sequence comprises m reference operation data, and in the distance matrix, the matrix value of the ith row and the jth column is the distance between the ith current operation data and the jth reference operation data; obtaining an accumulation matrix according to the distance matrix, wherein a matrix value in the accumulation matrix is a minimum accumulation distance, and the minimum accumulation distance is obtained according to the distance value in the distance matrix and a matrix value positioned in front in the accumulation matrix; and obtaining the distance between the current operation data sequence and the reference operation data sequence according to the matrix value in the accumulation matrix.

In some embodiments, the target device is a blanking device, the reference operating data sequence set is an operating data sequence corresponding to a device fault, and the acquiring a current operating data sequence corresponding to the target device includes: acquiring a current displacement data sequence corresponding to the blanking equipment, wherein the current displacement data sequence comprises current displacement data corresponding to the blanking equipment at a plurality of moments in a current time period; the comparing according to the target distance set and the distance threshold to obtain a comparison result, and the obtaining of the fault prediction result corresponding to the target device according to the comparison result includes: and when the distance in the target distance set is smaller than the distance threshold, determining that the corresponding fault prediction result of the blanking equipment is blocking.

In some embodiments, the obtaining of the distance threshold comprises: acquiring a standard operation data sequence set, wherein a fault operation data sequence in the standard operation data sequence set is an operation data sequence corresponding to an equipment fault; acquiring a normal operation data sequence set, and performing distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance; acquiring a fault operation data sequence set, and performing distance calculation according to the fault operation data sequence set and the standard operation data sequence set to obtain a second distance; and determining the distance threshold according to the first distance and the second distance, wherein the distance threshold is smaller than or equal to the first distance and larger than or equal to the second distance.

In some embodiments, the performing distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance includes: respectively calculating the distance between each normal operation data sequence and each standard operation data sequence in the standard operation data sequence set to obtain a distance set respectively corresponding to each normal operation data sequence; acquiring the minimum distance in the distance set corresponding to each normal operation data sequence to form a minimum distance set; and calculating the minimum distance in the minimum distance set to obtain a first distance.

An apparatus for device failure prediction, the apparatus comprising: a current running data sequence obtaining module, configured to obtain a current running data sequence corresponding to a target device, where the current running data sequence includes multiple running data corresponding to the target device in a current time period, and the running data in the running data sequence is sorted according to running time; the device comprises a reference operation data sequence set acquisition module, a data processing module and a data processing module, wherein the reference operation data sequence set acquisition module is used for acquiring a reference operation data sequence set, and the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal; a target distance set composing module, configured to respectively calculate distances between the current running data sequence and each reference running data sequence in the reference running data sequence set, and compose a target distance set; and the fault prediction result obtaining module is used for comparing the target distance set with the distance threshold value to obtain a comparison result and obtaining a fault prediction result corresponding to the target equipment according to the comparison result.

A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the device failure prediction method when executing the computer program.

A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned device failure prediction method.

According to the equipment fault prediction method, the device, the computer equipment and the storage medium, the current operation data sequence and the reference operation data sequence set corresponding to the target equipment are obtained, the distance between the current operation data sequence and each reference operation data sequence in the reference operation data sequence set is respectively calculated to form a target distance set, comparison is carried out according to the target distance set and the distance threshold value to obtain a comparison result, and the fault prediction result corresponding to the target equipment is obtained according to the comparison result, so that automatic fault prediction of the equipment is realized, and the efficiency of equipment fault prediction is improved.

Drawings

FIG. 1 is a diagram of an application scenario of a method for device failure prediction in some embodiments;

FIG. 2A is a schematic flow chart of a method for device failure prediction in some embodiments;

FIG. 2B is a flow diagram illustrating the operation of a data set in some embodiments;

FIG. 3 is a flow diagram illustrating a method for device failure prediction in some embodiments;

FIG. 4 is a flow diagram illustrating a method for device failure prediction in some embodiments;

FIG. 5A is a schematic flow chart illustrating the step of calculating the distance between the current operational data sequence and the reference operational data sequence in some embodiments;

FIG. 5B is a schematic diagram of a method for calculating a distance between operational data sequences according to some embodiments;

FIG. 5C is a schematic diagram of a method for calculating a distance between operational data sequences according to some embodiments;

FIG. 6 is a flow chart illustrating the distance threshold obtaining step in some embodiments;

FIG. 7 is a schematic flow chart of the first distance obtaining step in some embodiments;

FIG. 8 is a flow diagram illustrating a method for device failure prediction in some embodiments;

FIG. 9 is a block diagram of an apparatus for predicting device failure in some embodiments;

FIG. 10 is a diagram of the internal structure of a computer device in some embodiments.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The device failure prediction method provided by the application can be applied to the application environment shown in fig. 1. The application environment includes a target device 102, sensors within the target device, and a server 104. The target device 102 and the sensor may communicate with each other in a wired or wireless manner, and the sensor is used for acquiring operation data of the target device 102. The target device 102 communicates with the server 104 via a network through internal sensors via the network. The sensors may send the collected data to the server 104.

Specifically, the server 104 may obtain, through the sensor, a fault operation data sequence acquired within a period of time before the target device 102 fails and a normal operation data sequence within a period of time when the target device 102 normally operates, and obtain the distance threshold by calculating a distance between the operation data sequences according to the fault operation data sequence and the normal operation data sequence. The server 104 may obtain the current operating data sequence, respectively calculate the distance between the current operating data sequence and each reference operating data sequence in the reference operating data sequence set to form a target distance set, compare the target distance set with the calculated distance threshold to obtain a comparison result, and obtain a fault prediction result corresponding to the target device according to the comparison result. When the failure prediction result is a failure, the server 104 may feed back the failure prediction result to a relevant person, such as a maintenance person, or issue an apparatus stop instruction so that the apparatus may stop production. The target device 102 may be a production device, among others. The sensor may be one or more, and may include various types of sensors, such as a displacement sensor, an acceleration sensor, and the like, and the server 104 may be implemented by a separate server or a server cluster composed of a plurality of servers.

In some embodiments, as shown in fig. 2A, a method for predicting a device failure is provided, which is described by taking the method as an example applied to the server 104 in fig. 1, and includes the following steps:

s202, obtaining a current operation data sequence corresponding to the target device, wherein the current operation data sequence comprises a plurality of operation data corresponding to the target device in a current time period, and the operation data in the operation data sequence are sequenced according to operation time.

Specifically, the target device is a device that needs to perform failure prediction, and may be a production device, such as a feed opening. The operation data refers to data generated when the device operates, and may be at least one of displacement, velocity, or acceleration when the device operates, for example. The operation data sequence refers to a sequence composed of operation data, and the operation data in the operation data sequence may be sorted according to time. The current time period is the time period at which the current time is. The time length of the current time period corresponding to the operation data sequence may be set as needed, and the operation data corresponding to each time in the current time period may be obtained, for example, if the length of the current time period may be 10 minutes, one operation data corresponding to each minute may be obtained, and the current operation data sequence is formed. The data corresponding to each minute may be instantaneous operation data corresponding to a certain time, for example, 5 th second, or the operation data per second may be counted to obtain statistical data, for example, an average value of the operation data in the minute.

In some embodiments, the running data sequence is a time series (time series). The time series refers to a series formed by arranging numerical values of the same statistical index according to the time sequence of occurrence of the numerical values. The time in the time series may be in the form of a year, a quarter, a month, or any other form of time.

In some embodiments, the running data sequence is obtained by segmenting the obtained running data set according to a preset rule. The operation data in the operation data set may be sent by a sensor corresponding to the target device, and the operation data in the operation data set may be sorted according to time. The operation data set may be segmented according to a preset duration and a preset step length to obtain an operation data sequence, as shown in fig. 2B. Time in the first column of the table in fig. 2B represents time, value in the second column represents run data, and a set of data corresponding to the value column is a run data set. P1 and P2 are running data sequences obtained by dividing the running data set according to the duration T and the step size N. In general, the duration T is greater than the step size N.

S204, acquiring a reference operation data sequence set, wherein the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal.

Specifically, the reference operating data sequence set is a set composed of reference operating data sequences, and may be a historical operating data sequence corresponding to the device. For example, the reference operation data sequence may be an operation data sequence corresponding to a device failure, or may be an operation data sequence corresponding to a device normal. The running data sequence corresponding to the equipment fault is a sequence formed by corresponding running data within preset time before the equipment fault occurs, and the running data sequence corresponding to the normal equipment is a sequence formed by corresponding running data within preset time when the equipment runs normally. For example, when a device failure is determined, operational data within 10 minutes before the failure occurs may be acquired, making up an operational data sequence. The operation data sequence corresponding to the equipment fault can correspond to the fault type, the operation data sequence corresponding to the equipment fault is different, and the corresponding fault types can be the same or different.

In some embodiments, the set of constituent reference operational data sequences consists of corresponding operational data sequences prior to the equipment failure. Specifically, a plurality of operation data before the device failure may be collected, and a plurality of reference operation data sequences may be obtained according to the collected plurality of operation data corresponding to the failed device, so as to form a reference operation data sequence set.

And S206, respectively calculating the distance between the current running data sequence and each reference running data sequence in the reference running data sequence set to form a target distance set.

Specifically, the target distance set is a set formed by distances between the current operating data sequence and each reference operating data sequence in the reference operating data sequence set. For example, assuming that the current running data sequence is a, the reference running data sequence set includes B, C and D, the distance between a and B is AB, the distance between a and C is AC, and the distance between a and D is AD, the target distance set is a set consisting of the distance AB, the distance AC, and the distance AD. The distance calculation method may be set as needed, and for example, the euclidean distance method may be used for calculation.

In some embodiments, distance is used to reflect similarity, with smaller distances being more similar and larger distances being less similar. For example, for time series Q and time series C, the smaller the distance between time series Q and time series C, the greater the similarity.

In some embodiments, the distance between the running data sequences may be obtained according to a calculation method for calculating the Time series similarity, and the calculation method for calculating the Time series similarity may be, for example, a DTW (Dynamic Time Warping) algorithm. The DTW algorithm is based on the idea of dynamic programming, and can measure the similarity between two time sequences with inconsistent lengths.

And S208, comparing the target distance set with the distance threshold value to obtain a comparison result, and obtaining a fault prediction result corresponding to the target equipment according to the comparison result.

Specifically, the distance threshold may be preset, may be obtained empirically, or may be calculated by using a large amount of data. For example calculated over a large number of running data sequences. The failure prediction result may be a normal device or an abnormal device, where the abnormal device may be a specific abnormal device type, for example, a sudden shutdown of the device, etc.

In some embodiments, the minimum distance in the target distance set may be obtained, and the minimum distance in the target distance set is compared with a distance threshold to obtain a comparison result.

In some embodiments, an average value of distances in the target distance set may be calculated, and the average value of the distances in the target distance set is compared with a distance threshold to obtain a comparison result.

In some embodiments, when the comparison result is smaller than the distance threshold, the failure prediction result corresponding to the target device is determined to be a device failure.

In some embodiments, when the comparison result is greater than the distance threshold, the failure prediction result corresponding to the target device is determined to be a device failure.

In some embodiments, if the reference operating data sequence set includes an operating data sequence normally corresponding to the device, and the distance threshold is a distance threshold normally corresponding to the device, the prediction result is that the device is normal when the target distance is smaller than the distance threshold.

In some embodiments, if the reference operating data sequence set includes an operating data sequence corresponding to an equipment fault and the distance threshold is a distance threshold corresponding to the equipment fault, the prediction result is the equipment fault when the target distance is less than the distance threshold.

For example, the DTW operation is performed on the time sequence to be measured and the time sequence with the fault to obtain a value of 1.02, the DTW operation is performed on the time sequence with the non-fault to obtain a value of 9.24, and if the distance threshold corresponding to the equipment fault is 2. The time series to be compared is considered as a fault time series, and the feed opening is predicted to be blocked.

According to the equipment fault prediction method, the current operation data sequence and the reference operation data sequence set corresponding to the target equipment are obtained, the distance between the current operation data sequence and each reference operation data sequence in the reference operation data sequence set is calculated respectively to form a target distance set, comparison is carried out according to the target distance set and a distance threshold value to obtain a comparison result, and the reference operation data sequence is a historical operation data sequence, so that future prediction can be carried out, a fault prediction result corresponding to the target equipment can be obtained according to the comparison result, automatic fault prediction of the equipment is achieved, and the efficiency of equipment fault prediction is improved.

The equipment fault prediction method can be used for predicting the fault of the equipment in the Internet of things equipment. At the present stage, the fault prediction of the industrial internet-of-things equipment is still in a starting stage, for example, the blockage prediction of a feed opening needs to be carried out by means of handheld vibration fault detection equipment, and great inconvenience is brought to enterprises. In addition, because the technology of Internet of things (IOT) and big data (big data) is not used, the operation data collected by the handheld vibration fault detection device does not have strong generalization capability and cannot be popularized greatly, and the handheld vibration fault detection device needs a large amount of manpower for maintenance, so that the device fault prediction efficiency is low. By adopting the equipment fault prediction method, the fault is predicted by acquiring the corresponding operation data sequence before the equipment sends the fault. Automatic fault prediction of equipment in the Internet of things can be achieved, and therefore the efficiency of the fault prediction of the equipment in the Internet of things is improved.

The internet of things, namely the internet connected with everything, is an extended and expanded network on the basis of the internet, is a huge network formed by combining various information sensing devices and the internet, and realizes interconnection and intercommunication of people, machines and things at any time and any place. Big data is a data set which cannot be captured, managed and processed by a conventional software tool within a certain time range, and is a massive, high-growth-rate and diversified information asset which can have stronger decision-making power, insight discovery power and flow optimization capability only by a new processing mode.

In some embodiments, referring to the operation data sequence set as the operation data sequence corresponding to the device fault, as shown in fig. 3, the step S208 compares the target distance set with the distance threshold to obtain a comparison result, and obtains a fault prediction result corresponding to the target device according to the comparison result, including:

s302, calculating the minimum distance in the target distance set.

S304, comparing the minimum distance with a distance threshold value to obtain a comparison result.

S306, when the comparison result shows that the minimum distance is smaller than the distance threshold, determining that the fault prediction result corresponding to the target device is a fault.

Specifically, the target distance set includes a plurality of distances, the comparison result may be a difference between the minimum distance and a distance threshold, the minimum distance is smaller than the distance threshold when the comparison result is smaller than zero, and the minimum distance is larger than the distance threshold when the comparison result is larger than zero. The server may calculate a minimum distance in the target distance set, compare a size relationship between the minimum distance and a distance threshold, and determine that a failure prediction result corresponding to the target device is a failure when the minimum distance is smaller than the distance threshold.

In the embodiment of the present application, by comparing the minimum distance in the target distance set with the distance threshold, since the reference operating data sequence set is the operating data sequence corresponding to the equipment fault, and the similarity between the reference operating data sequence corresponding to the minimum distance and the current operating data sequence is the maximum, the accuracy of the obtained fault prediction result is high by comparing the minimum distance with the distance threshold.

In some embodiments, the failure prediction result further includes a predicted failure type, as shown in fig. 4, when the comparison result is that the minimum distance is smaller than the distance threshold, determining that the failure prediction result corresponding to the target device is a failure in step S306 includes:

s402, when the comparison result shows that the minimum distance is smaller than the distance threshold, acquiring a reference running data sequence corresponding to the minimum distance as a target reference running data sequence.

S404, acquiring a fault type corresponding to the target reference operation data sequence as a predicted fault type corresponding to the target equipment.

Specifically, there may be a plurality of reference operating data sequences, and different reference operating data sequences may correspond to different fault types or the same fault type. For example, for the reference operation data sequence S1, the corresponding fault type may be E1, for the reference operation data sequence S2, the corresponding fault type may be E2, and for the reference operation data sequence S3, the corresponding fault type may also be E2. When the comparison result is determined that the minimum distance is smaller than the distance threshold, the reference operation data sequence with the minimum distance can be obtained, and then the corresponding fault type can be obtained. The types of failures may be, for example, equipment shutdown, equipment jam, equipment outage, and the like.

In some embodiments, the reference operation data sequence and the corresponding fault type are stored in association in the server, and the corresponding fault type may be obtained according to the reference operation data sequence.

In the embodiment of the application, according to the minimum distance, the reference operation data sequence corresponding to the minimum distance is obtained, and the fault type corresponding to the reference operation data sequence corresponding to the minimum distance is obtained, so that the type of the fault occurring in the target device can be obtained, and therefore, the efficiency of performing problem location and problem solving on the device by a maintenance worker can be improved.

In some embodiments, as shown in fig. 5A, the step S206 of calculating the distance between the current running data sequence and each reference running data sequence in the reference running data sequence set respectively includes:

s502, a distance matrix is obtained according to the current operation data sequence and the reference operation data sequence, the current operation data sequence comprises n current operation data, the reference operation data sequence comprises m reference operation data, and in the distance matrix, the matrix value of the ith row and the jth column is the distance between the ith current operation data and the jth reference operation data.

In particular, the distance matrixThe matrix value in the ith row and the jth column may be an absolute value of a difference between the ith current operation data and the jth reference operation data, or may be a square of a difference between the ith current operation data and the jth reference operation data. As shown in fig. 5B, a distance matrix corresponding to time series a and time series B is shown, where the length of time series a is n, the length of time series B is m, and a ═ q is assumed to be q₁,q₂,...,q_i,...,q_n，B＝c₁,c₂,...,c_j,...,c_m. The matrix element corresponding to position (i, j) in the distance matrix represents q_iAnd c_jDistance between two points, i.e. q_iAnd c_jSimilarity between two points. q. q.s_iAnd c_jThe distance between two points can be d (q)_i,c_j) Is represented by d (q)_i,c_j) Smaller is q_iAnd c_jThe higher the similarity between two points. The distance (similarity) between two data is generally calculated using Euclidean distances, such as d (q)_i,c_j)＝(q_i-c_j)²The absolute value of the data difference may also be used as the distance between the data.

In some embodiments, the distance matrix may be n × m or m × n. For example, for a current running data sequence q of length 6 ═ 3, 5,6, 7, 7, 1], a reference running data sequence c of length 7 ═ 3, 6, 6,7, 8, 1, 1], the corresponding distance matrix may be 6 × 7 or 7 × 6. Wherein the 6 x 7 distance matrix is shown in table 1. Each of [3, 5,6, 7, 7, 1] is operation data. The bold content in table 1 is the reference run data sequence [3, 6, 6,7, 8, 1, 1], and the italic content is the current run data sequence [3, 5,6, 7, 7, 1 ]. The content of the grey area represents the distance matrix, which can be seen to be 6 rows and 7 columns. The position of the reference operation data sequence and the current operation data sequence in table 1 is exchanged to obtain a 7 × 6 distance matrix. The distance value in the distance matrix is the absolute value of the difference between the current operation data and the reference operation data, the matrix value in the 1 st row and the 1 st column is the distance between the 1 st current operation data and the 1 st reference operation data, that is, the matrix value in the 1 st row and the 1 st column is the absolute value of the difference between 3 and 3, and the matrix value in the 1 st row and the 1 st column in the table is 0 because the difference between 3 and 3 is 0.

TABLE 1

In some embodiments, if n ═ m, the distance between time series a and time series B may be calculated directly; if n is not equal to m, time series A and time series B may be aligned first. Of course, if n ═ m, the alignment operation may be performed. To align time series a and time series B, an n × m matrix grid may be constructed.

S504, an accumulation matrix is obtained according to the distance matrix, a matrix value in the accumulation matrix is a minimum accumulation distance, and the minimum accumulation distance is obtained according to the distance value in the distance matrix and a matrix value positioned in front in the accumulation matrix.

Specifically, a corresponding accumulation matrix may be obtained according to the distance matrix, and a distance between the current operation data sequence and the reference operation data sequence may be obtained according to the accumulation matrix. The matrix value in the accumulation matrix is the minimum accumulation distance from the starting point to the point, and the minimum accumulation distance can be calculated according to the formula (1).

D(i,j)＝d(q_i,c_j)+min{D(i-1,j-1),D(i-1,j),D(i,j-1)} (1)

Where D (i, j) represents the minimum accumulated distance (cumulative distance) from the start position of the distance matrix to position (i, j), as shown in FIG. 5C, D (i, j) can only be obtained from the three minimum accumulated distances D (i-1, j), D (i, j-1) and D (i-1, j-1). Equation (1) indicates that the minimum accumulated distance D (i, j) is the distance value D (q) corresponding to the position of (i, j)_i,c_j) The sum of the smallest accumulated distances corresponding to the smallest neighboring elements that can reach the (i, j) position. The smaller D (i, j), the higher the similarity of the data. For example, the accumulation matrix corresponding to table 1 can be obtained according to equation (1), as shown by the gray area in table 2. Taking the cumulative distance 1 in row 2 and column 2 as an example, the minimum cumulative distance in row 2 and column 2 can be represented as D (2,2), and can be determined only from three of D (1,2), D (2,1) and D (1,1)The minimum value among the cumulative distances D (1,2) ═ 3, D (2,1) ═ 2, and D (1,1) ═ 0, and thus the minimum value among the three cumulative distances D (1,2), D (2,1), and D (1,1) is D (1,1) ═ 0, and thus D (2,2) ═ D (q) (q ═ 0)₂,c₂) D (1,1) ═ D (5,6) + D (1,1), where D (5,6) denotes the distance between data 5 and data 6, and the absolute value of data 5 and data 6 can be taken as the distance between data 5 and data 6, so D (5,6) ═ 1, and thus D (2,2) ═ 1+0 = 1.

TABLE 2

S506, according to the matrix value in the accumulation matrix, the distance between the current operation data sequence and the reference operation data sequence is obtained.

Specifically, a matrix value of an end position in the accumulation matrix may be obtained as a distance between the current running data sequence and the reference running data sequence, and for the accumulation matrix of n rows and m columns, the end position refers to an nth row and an mth column. For example, it can be obtained from table 2 that D (6,7) is 1, that is, the minimum distance corresponding to the shortest path from the start point to the end point of the matrix is 1, and therefore, the distance between the current run data sequence q and the reference run data sequence c is 1.

In some embodiments, the distance between the current running data sequence and the reference running data sequence corresponds to a shortest path, where the shortest path is a position corresponding to a matrix value involved in obtaining a matrix value with a value as an end point in the accumulation matrix. For example, the result D (6,7) relates to the dark gray area in table 2, and thus, the shortest path corresponding to the distance between the current running data sequence q and the reference running data sequence c is the portion of the dark gray area in table 2. Therefore, the distance value corresponding to the shortest path from the starting point to the end point of the matrix is obtained, and the distance between the current operation data sequence and the reference operation sequence can be accurately reflected through the distance value of the shortest path, so that the accuracy of calculating the distance between the current operation data sequence and the reference operation data sequence is improved.

In the embodiment of the present application, the corresponding accumulation matrix is obtained according to the matrix value in the distance matrix, and the distance between the current running data sequence and the reference running data sequence is obtained according to the matrix value in the accumulation matrix.

In some embodiments, the target device is a blanking device, the reference operation data sequence set is an operation data sequence corresponding to a device fault, and the acquiring of the current operation data sequence corresponding to the target device in step S202 includes: and acquiring a current displacement data sequence corresponding to the blanking equipment, wherein the current displacement data sequence comprises current displacement data respectively corresponding to the blanking equipment at a plurality of moments in a current time period. Step S208, comparing the target distance set with the distance threshold value to obtain a comparison result, wherein the step of obtaining the fault prediction result corresponding to the target equipment according to the comparison result comprises the following steps: and when the distance in the target distance set is smaller than the distance threshold, determining that the corresponding fault prediction result of the blanking equipment is blocking.

In particular, the blanking device is used for transporting material, which can be output from the blanking device. The current displacement data sequence is a sequence formed by displacement data of the blanking equipment in the current time period during vibration. The distance threshold value can be obtained according to a corresponding displacement data sequence when the blanking device is blocked and a corresponding displacement data sequence when the blanking device is not blocked, and when the minimum distance is smaller than the distance threshold value, the current displacement data sequence can be considered as the corresponding displacement data sequence when the blanking device is blocked, so that the corresponding failure prediction result of the blanking device can be determined as the blockage.

In some embodiments, when there is a distance smaller than the distance threshold in the target distance set, it may be determined that the failure prediction result corresponding to the blanking device is a jam.

In some embodiments, there may be a plurality of distances smaller than the distance threshold in the target distance set, and when there are a preset number of distances smaller than the distance threshold in the target distance set, it may be determined that the failure prediction result corresponding to the blanking device is a jam.

In the embodiment of the application, the minimum distance is obtained according to the current displacement data sequence corresponding to the blanking equipment and the operation data sequence corresponding to the equipment fault, the minimum distance is compared with the distance threshold to obtain the comparison result, and when the comparison result is that the minimum distance is smaller than the distance threshold, the fault prediction result corresponding to the blanking equipment is determined to be blocking.

In some embodiments, the distance threshold is calculated in advance by acquiring historical operating data, and as shown in fig. 6, the step of obtaining the distance threshold includes:

s602, a standard operation data sequence set is obtained, and a fault operation data sequence in the standard operation data sequence set is an operation data sequence corresponding to the equipment fault.

S604, acquiring a normal operation data sequence set, and performing distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance.

S606, acquiring a fault operation data sequence set, and performing distance calculation according to the fault operation data sequence set and the standard operation data sequence set to obtain a second distance.

S608, according to the first distance and the second distance, a distance threshold is determined, and the distance threshold is smaller than or equal to the first distance and larger than or equal to the second distance.

Specifically, the failure operation data sequence set is a set composed of operation data sequences corresponding to equipment failures. The plurality of fault operation data sequences obtained in advance can be grouped to obtain a standard operation data sequence set and a fault operation data sequence set, and the operation data sequences normally corresponding to the equipment obtained in advance can be combined into a normal operation data sequence set.

In some embodiments, the distances between each normal operation data sequence and each standard operation data sequence in the standard operation data sequence set may be calculated, and the first distance may be obtained according to the calculated distances.

In some embodiments, the distance between each faulty operating data sequence and each standard operating data sequence in the set of standard operating data sequences may be calculated, and the second distance may be obtained according to the calculated distance.

In some embodiments, a region having a maximum value of the first distance and a minimum value of the second distance may be obtained, and data is selected in the region as the distance threshold.

In the embodiment of the present application, a distance is calculated according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance, a distance is calculated according to the fault operation data sequence set and the standard operation data sequence set to obtain a second distance, and the standard operation data sequence set includes the fault operation data sequence, so that the first distance may represent a maximum distance between the normal operation data sequence and the fault operation data sequence in the standard operation data sequence set, and the second distance may represent a maximum distance between the fault operation data sequence and the fault operation data sequence in the standard operation data sequence set, and therefore, an accuracy of identifying the normal operation data sequence and the fault operation data sequence according to a distance threshold determined by the first distance and the second distance is high.

In some embodiments, as shown in fig. 7, performing distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain the first distance includes:

s702, respectively calculating the distance between each normal operation data sequence and each standard operation data sequence in the standard operation data sequence set to obtain a distance set respectively corresponding to each normal operation data sequence.

Specifically, if the normal operation data sequence includes A1 and A2, the standard operation data sequence includes B1, B2 and B3, and the distances between A1 and B1, B2 and B3 are A1B1, A1B2 and A1B3, respectively, and the distances between A2 and B1, between B2 and B3 are A2B1, A2B2 and A2B3, respectively. The distance set corresponding to the normal operation data sequence A1 is { A1B1, A1B2, A1B3}, and the distance set corresponding to the normal operation data sequence A2 is { A2B1, A2B2, A2B3 }.

S704, obtaining the minimum distance in the distance set corresponding to each normal operation data sequence to form a minimum distance set.

Specifically, the minimum distance is the distance with the smallest distance. Specifically, if the distance set corresponding to the normal operation data sequence A1 is a minimum distance A1B1 in { A1B1, A1B2, A1B3}, and the distance set corresponding to the normal operation data sequence A2 is a minimum distance A2B1 in { A2B1, A2B2, A2B3}, the minimum distance set is { A1B1, A2B1 }.

S706, calculating the minimum distance in the minimum distance set to obtain a first distance.

Specifically, if the minimum distance set includes a first minimum distance, a second minimum distance, and a third minimum distance, and the first minimum distance is the smallest of the three, the first minimum distance is taken as the first distance. For example, if the minimum distance set is { A1B1, A2B1}, and A1B1 is the smallest of A1B1 and A2B1, then A1B1 is taken as the first distance.

In some embodiments, obtaining the fault operation data sequence set, and performing distance calculation according to the fault operation data sequence set and the standard operation data sequence set to obtain the second distance includes: respectively calculating the distance between each fault operation data sequence and each standard operation data sequence in the standard operation data sequence set to obtain a distance set corresponding to each fault operation data sequence; acquiring minimum distances in distance sets respectively corresponding to all fault operation data sequences to form a first minimum distance set; and calculating the maximum distance in the first minimum distance set to obtain a second distance.

In the embodiment of the present application, the minimum distance in the minimum distance set is calculated, and the minimum distance in the minimum distance set is used as the first distance, because the minimum distance in the minimum distance set can reflect the minimum distance between the normal operation data sequence and the fault operation data sequence in the standard operation data sequence set, the accuracy of the obtained first distance is high.

In some embodiments, a flowchart corresponding to the device failure prediction method of the present application is described in fig. 8. The flow in fig. 8 is accomplished through a large data platform. The data (data) in fig. 8 may be vibration data of the blanking device sent by the vibration sensor, the vibration data is related to time, the vibration data may be, for example, displacement data of a certain time, velocity data of a certain time, acceleration data of a certain time, or the like, and the large data platform may sort the plurality of vibration data by time to obtain a time sequence. The vibration sensor can send data to the big data platform through an LoRa protocol. Of course, the vibration sensor may also transmit data to the big data platform by a WiFi (Wireless Fidelity) technology. Fig. 8 may include a first phase and a second phase. The first stage is a model training stage, and the distance threshold is obtained through model training. And the second stage is a stage of predicting the fault of the equipment in real time according to the distance threshold value. Wherein the first stage may comprise:

step 1, data receiving and storing.

Specifically, vibration sensor gathers unloading equipment's vibration data, sends vibration data for big data platform through the loRa protocol. The big data analysis platform receives the operation data through the kafka component, and stores the operation data in a local data warehouse through the kafka component, wherein the kafka component is a high-throughput Distributed publish-subscribe message System, and the local data warehouse can be a big data storage layer and can include hdfs (Hadoop Distributed File System), hbase and hive. The hbase is a distributed and column-oriented open source database, and the hive is a data warehouse tool used for data extraction, conversion and loading.

And 2, preprocessing vibration data.

Specifically, ETL (Extract-Transform-Load) is performed on the data. In which ETL is used to describe the process of extracting (extract), converting (transform), and loading (load) data from a source to a destination. The big data platform can extract, clean and convert vibration data through a Spark SQL component and then load the vibration data into a data warehouse, and mainly comprises filling of data missing, and removing and correcting of abnormal values. Wherein, the Spark SQL component is a Spark offline processing component.

And step 3: the data slices are vibrated.

Specifically, the vibration data is time-related data, and the sequence of the vibration data is a time sequence. The vibration data may be sliced to obtain a plurality of vibration data sequences. The big data platform can use Spark SQL to divide the vibration data into time series of the same size according to time. Time series based data S_rawCan be cut into a plurality of time segments P according to a time window of K and a step length of L_i＝[t_i,S_i]Wherein each time slice comprises all sensor data information comprised within the time window. Wherein i is 1,2,3 …. The vibration data may be data when the blanking apparatus is normal, or may be data when the blanking apparatus is out of order. The non-failure time sequence can be obtained according to the data of the blanking equipment when the blanking equipment is normal, and the failure time sequence can be obtained according to the data of the blanking equipment when the blanking equipment fails. The time window is typically a multiple of the time series data period, the step size is also a multiple of the time series data period, and the time window is preferably greater than the length of the step size. The time series data period is not necessarily the same as the data period determined by the equipment, different machines and different machine characteristics, so that different equipment can correspond to different time windows K and step lengths L.

And 4, step 4: a distance threshold is calculated.

Specifically, when a plurality of failure time series and non-failure time series are obtained, the plurality of failure time series may be grouped into two sets, a first set and a second set, and the plurality of non-failure time series may be grouped into a third set. A failure time series may be selected from the first group and each failure time series in the second group may be respectively subjected to distance calculation to obtain a minimum value of the distance, which is used as a first maximum value, a non-failure time series may be selected from the third group and each failure time series in the second group may be respectively subjected to distance calculation to obtain a minimum value of the distance, which is used as a second maximum value, and the distance threshold may be defined between the first maximum value and the second maximum value (which may include at least one of the first maximum value or the second maximum value). And repeating the process, continuously adjusting the distance threshold value to obtain the optimal distance threshold value, and deploying the optimal distance threshold value to the application layer.

The second stage may include:

step 1, continuously receiving and processing data in Kafka through Spark streaming (big data Spark frame real-time processing component) to obtain a time sequence to be detected;

and 2, loading the time sequence to be tested into a model with a distance threshold value, and predicting the time sequence to be tested.

Above-mentioned this application embodiment adopts the vibration sensor to gather thing networking device's vibration data to transmit data to big data analysis platform through loRa agreement, receive the time sequence that obtains the vibration data and constitute through big data analysis platform analysis, through the DTW technique, establish the failure prediction model, thereby the emergence of accurate prediction feed opening trouble. Therefore, the machine learning model based on the vibration signals is established by analyzing a large number of time sequences based on the data collected by the Internet of things and through the big data analysis capability, the blockage situation of the feed opening is predicted, and the generalization capability of the model is greatly improved.

It should be understood that, although the steps in the flowcharts of the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts of the above embodiments may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or the stages is not necessarily sequential, but may be performed alternately or alternatingly with other steps or at least a part of the sub-steps or the stages of other steps.

In some embodiments, as shown in fig. 9, there is provided an apparatus for predicting a device failure, including: a current operation data sequence obtaining module 902, a reference operation data sequence set obtaining module 904, a target distance set composing module 906, and a failure prediction result obtaining module 908, wherein:

a current running data sequence obtaining module 902, configured to obtain a current running data sequence corresponding to the target device, where the current running data sequence includes multiple running data corresponding to the target device in a current time period, and the running data in the running data sequence is sorted according to running time.

A reference operating data sequence set obtaining module 904, configured to obtain a reference operating data sequence set, where the reference operating data sequence set includes at least one of an operating data sequence corresponding to a device failure or an operating data sequence corresponding to a device normal.

A target distance set composing module 906, configured to respectively calculate distances between the current running data sequence and each reference running data sequence in the reference running data sequence set, so as to compose a target distance set.

A failure prediction result obtaining module 908, configured to compare the target distance set with the distance threshold to obtain a comparison result, and obtain a failure prediction result corresponding to the target device according to the comparison result.

In some embodiments, the reference operation data sequence set is an operation data sequence corresponding to the equipment fault, and the fault prediction result obtaining module 908 includes:

and the minimum distance calculation unit is used for calculating the minimum distance in the target distance set.

And the comparison result obtaining unit is used for comparing the minimum distance with the distance threshold value to obtain a comparison result.

And the failure prediction result determining unit is used for determining that the failure prediction result corresponding to the target equipment is a failure when the comparison result shows that the minimum distance is smaller than the distance threshold.

In some embodiments, the fault prediction result further includes a predicted fault type, and the fault prediction result determining unit is further configured to, when the comparison result is that the minimum distance is smaller than the distance threshold, obtain a reference operating data sequence corresponding to the minimum distance as a target reference operating data sequence; and acquiring a fault type corresponding to the target reference operation data sequence as a predicted fault type corresponding to the target equipment.

In some embodiments, the set of target distances constitutes module 906, including:

and the distance matrix obtaining unit is used for obtaining a distance matrix according to the current operation data sequence and the reference operation data sequence, the current operation data sequence comprises n current operation data, the reference operation data sequence comprises m reference operation data, and in the distance matrix, the matrix value of the ith row and the jth column is the distance between the ith current operation data and the jth reference operation data.

And the accumulation matrix obtaining unit is used for obtaining an accumulation matrix according to the distance matrix, wherein a matrix value in the accumulation matrix is a minimum accumulation distance, and the minimum accumulation distance is obtained according to the distance value in the distance matrix and a matrix value positioned in front in the accumulation matrix.

And the distance obtaining unit is used for obtaining the distance between the current operation data sequence and the reference operation data sequence according to the matrix value in the accumulation matrix.

In some embodiments, the target device is a blanking device, the reference operation data sequence set is an operation data sequence corresponding to a device fault, and the current operation data sequence obtaining module 902 is further configured to obtain a current displacement data sequence corresponding to the blanking device, where the current displacement data sequence includes current displacement data corresponding to the blanking device at multiple times in a current time period; the failure prediction result obtaining module 908 is further configured to determine that the failure prediction result corresponding to the blanking device is blocked when the distance in the target distance set is smaller than the distance threshold.

In some embodiments, the device failure prediction apparatus further includes a distance threshold obtaining module, where the distance threshold obtaining module includes:

and the standard operation data sequence set acquisition unit is used for acquiring a standard operation data sequence set, and the fault operation data sequence in the standard operation data sequence set is an operation data sequence corresponding to the equipment fault.

And the first distance obtaining unit is used for obtaining the normal operation data sequence set and carrying out distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance.

And the second distance obtaining unit is used for obtaining the fault operation data sequence set and carrying out distance calculation according to the fault operation data sequence set and the standard operation data sequence set to obtain a second distance.

And the distance threshold value determining unit is used for determining a distance threshold value according to the first distance and the second distance, wherein the distance threshold value is smaller than or equal to the first distance and larger than or equal to the second distance.

In some embodiments, the first distance obtaining unit is further configured to calculate distances between each normal operation data sequence and each standard operation data sequence in the standard operation data sequence set, respectively, to obtain a distance set corresponding to each normal operation data sequence; acquiring the minimum distance in the distance set corresponding to each normal operation data sequence to form a minimum distance set; and calculating the minimum distance in the minimum distance set to obtain the first distance.

For specific limitations of the device failure prediction apparatus, see the above limitations on the device failure prediction method, which are not described herein again. The modules in the device failure prediction apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In some embodiments, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a device failure prediction method.

Those skilled in the art will appreciate that the architecture shown in fig. 10 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In some embodiments, a computer device is provided, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the device failure prediction method when executing the computer program.

In some embodiments, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the above-mentioned device failure prediction method.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method of device failure prediction, the method comprising:

acquiring a current operation data sequence corresponding to target equipment, wherein the current operation data sequence comprises a plurality of operation data corresponding to the target equipment in a current time period, and the operation data in the operation data sequence are sequenced according to operation time;

acquiring a reference operation data sequence set, wherein the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal;

respectively calculating the distance between the current operating data sequence and each reference operating data sequence in the reference operating data sequence set to form a target distance set;

and comparing the target distance set with a distance threshold value to obtain a comparison result, and obtaining a fault prediction result corresponding to the target equipment according to the comparison result.

2. The method according to claim 1, wherein the step of obtaining the reference operating data sequence set as an operating data sequence corresponding to the device fault by comparing the target distance set with a distance threshold to obtain a comparison result, and obtaining the fault prediction result corresponding to the target device according to the comparison result comprises:

calculating a minimum distance in the set of target distances;

comparing the minimum distance with the distance threshold value to obtain a comparison result;

and when the comparison result is that the minimum distance is smaller than the distance threshold, determining that the fault prediction result corresponding to the target equipment is a fault.

3. The method according to claim 2, wherein the fault prediction result further includes a predicted fault type, and when the comparison result indicates that the minimum distance is smaller than the distance threshold, determining that the fault prediction result corresponding to the target device is a fault includes:

when the comparison result is that the minimum distance is smaller than the distance threshold, acquiring a reference operation data sequence corresponding to the minimum distance as a target reference operation data sequence;

and acquiring a fault type corresponding to the target reference operation data sequence as a predicted fault type corresponding to the target equipment.

4. The method of claim 1, wherein separately calculating the distance between the current operational data sequence and each reference operational data sequence in the set of reference operational data sequences comprises:

obtaining a distance matrix according to the current operation data sequence and the reference operation data sequence, wherein the current operation data sequence comprises n current operation data, the reference operation data sequence comprises m reference operation data, and in the distance matrix, the matrix value of the ith row and the jth column is the distance between the ith current operation data and the jth reference operation data;

obtaining an accumulation matrix according to the distance matrix, wherein a matrix value in the accumulation matrix is a minimum accumulation distance, and the minimum accumulation distance is obtained according to the distance value in the distance matrix and a matrix value positioned in front in the accumulation matrix;

and obtaining the distance between the current operation data sequence and the reference operation data sequence according to the matrix value in the accumulation matrix.

5. The method according to claim 1, wherein the target device is a blanking device, the reference operating data sequence set is an operating data sequence corresponding to a device fault, and the acquiring a current operating data sequence corresponding to the target device includes:

acquiring a current displacement data sequence corresponding to the blanking equipment, wherein the current displacement data sequence comprises current displacement data corresponding to the blanking equipment at a plurality of moments in a current time period;

the comparing according to the target distance set and the distance threshold to obtain a comparison result, and the obtaining of the fault prediction result corresponding to the target device according to the comparison result includes:

and when the distance in the target distance set is smaller than the distance threshold, determining that the corresponding fault prediction result of the blanking equipment is blocking.

6. The method of claim 1, wherein the step of deriving the distance threshold comprises:

acquiring a standard operation data sequence set, wherein a fault operation data sequence in the standard operation data sequence set is an operation data sequence corresponding to an equipment fault;

acquiring a normal operation data sequence set, and performing distance calculation according to the normal operation data sequence set and the standard operation data sequence set to obtain a first distance;

acquiring a fault operation data sequence set, and performing distance calculation according to the fault operation data sequence set and the standard operation data sequence set to obtain a second distance;

and determining the distance threshold according to the first distance and the second distance, wherein the distance threshold is smaller than or equal to the first distance and larger than or equal to the second distance.

7. The method of claim 6, wherein said performing a distance calculation based on said set of normal operating data sequences and said set of standard operating data sequences to obtain a first distance comprises:

respectively calculating the distance between each normal operation data sequence and each standard operation data sequence in the standard operation data sequence set to obtain a distance set respectively corresponding to each normal operation data sequence;

acquiring the minimum distance in the distance set corresponding to each normal operation data sequence to form a minimum distance set;

and calculating the minimum distance in the minimum distance set to obtain a first distance.

8. An apparatus for predicting a failure of a device, the apparatus comprising:

a current running data sequence obtaining module, configured to obtain a current running data sequence corresponding to a target device, where the current running data sequence includes multiple running data corresponding to the target device in a current time period, and the running data in the running data sequence is sorted according to running time;

the device comprises a reference operation data sequence set acquisition module, a data processing module and a data processing module, wherein the reference operation data sequence set acquisition module is used for acquiring a reference operation data sequence set, and the reference operation data sequence set comprises at least one of an operation data sequence corresponding to equipment failure or an operation data sequence corresponding to equipment normal;

a target distance set composing module, configured to respectively calculate distances between the current running data sequence and each reference running data sequence in the reference running data sequence set, and compose a target distance set;

and the fault prediction result obtaining module is used for comparing the target distance set with the distance threshold value to obtain a comparison result and obtaining a fault prediction result corresponding to the target equipment according to the comparison result.

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the device failure prediction method of any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the device failure prediction method of any one of claims 1 to 7.