CN117473312A

CN117473312A - Bearing state prediction method, bearing state prediction device, computer equipment and storage medium

Info

Publication number: CN117473312A
Application number: CN202311417796.6A
Authority: CN
Inventors: 蒋陈昱; 雷柏茂; 陈强; 李骞
Original assignee: China Electronic Product Reliability and Environmental Testing Research Institute
Current assignee: China Electronic Product Reliability and Environmental Testing Research Institute
Priority date: 2023-10-27
Filing date: 2023-10-27
Publication date: 2024-01-30

Abstract

The application relates to a bearing state prediction method, a bearing state prediction device, computer equipment and a storage medium, and relates to the field of nuclear engineering. The method comprises the following steps: determining the service cycle of a bearing to be detected, and determining a reference characteristic value of the bearing to be detected in the service cycle; according to the reference characteristic value, carrying out parameter updating on the initial characteristic change function to obtain a target characteristic change function; predicting a predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function; and carrying out state prediction on the predicted characteristic value by adopting a target characteristic change function and a hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period. The method and the device can conduct real-time running state analysis and prediction in bearing running.

Description

Bearing state prediction method, bearing state prediction device, computer equipment and storage medium

Technical Field

The present disclosure relates to the technical field and field of industrial equipment, and in particular, to a method and apparatus for predicting a bearing state, a computer device, and a storage medium.

Background

The bearing is used as a basic part with very wide application, and plays a role in supporting mechanical rotation in the operation process of a plurality of industrial equipment, so the reliability of the bearing is important for guaranteeing the safe and stable operation of the industrial equipment, and the reliability analysis and prediction of the bearing are also important points of attention in the technical field of industrial equipment.

In the prior art, a hidden Markov model can be trained through a large number of tests and fault data, and further, reliability prediction of a future time period is performed on a bearing according to the hidden Markov model, however, for many emerging industrial fields (such as large-scale wind turbines), the tests and fault data of bearing operation are less, so that the reliability prediction of the future time period cannot be accurately performed on the bearing according to the hidden Markov model.

Disclosure of Invention

In view of the above, it is desirable to provide a bearing state prediction method, apparatus, computer device, and storage medium that can accurately analyze and predict the reliability of a bearing.

In a first aspect, the present application provides a method of predicting a bearing condition. The method comprises the following steps:

determining the service cycle of the bearing to be detected, and determining the reference characteristic value of the bearing to be detected in the service cycle;

according to the reference characteristic value, carrying out parameter updating on the initial characteristic change function to obtain a target characteristic change function;

predicting a predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function;

and carrying out state prediction on the predicted characteristic value by adopting a target characteristic change function and a hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period.

In one embodiment, a state prediction is performed on a predicted feature value by using a target feature change function and a hidden markov model to obtain a predicted running state of a bearing to be detected in a future time period, including:

determining a degradation trend function according to the target characteristic change function;

updating the hidden Markov model according to the degradation trend function to obtain a hidden Markov degradation model;

and carrying out state prediction on the prediction characteristic value by adopting a hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

In one embodiment, a hidden markov degradation model is used to predict a state of a predicted feature value to obtain a predicted running state of a bearing to be detected in a future time period, including:

determining an initial running state of the bearing to be detected in a future time period, wherein the initial running state refers to a running state of the bearing to be detected corresponding to a previous time period of the future time period;

and carrying out state prediction on the prediction characteristic value according to the initial running state and the hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

In one embodiment, according to an initial running state and a hidden markov degradation model, performing state prediction on a prediction feature value to obtain a predicted running state of a bearing to be detected in a future time period, including:

Inputting the predicted characteristic value into a hidden Markov degradation model to obtain a candidate running state output by the hidden Markov degradation model;

a predicted operating state corresponding to the initial operating state is determined from the candidate operating states.

In one embodiment, the training process of the hidden Markov model includes:

determining a historical characteristic value of the sample bearing in a historical period, and determining a historical running state corresponding to the historical characteristic value;

and carrying out parameter adjustment on the state transition probability matrix parameters and the observation value probability matrix parameters of the initial model according to the historical characteristic values and the historical running states corresponding to the historical characteristic values to obtain the hidden Markov model.

In one embodiment, the method further comprises:

determining a historical characteristic value of the sample bearing in a historical period;

and determining characteristic change parameters corresponding to the bearing to be detected according to the historical characteristic values, and constructing an initial characteristic change function according to the characteristic change parameters.

In a second aspect, the present application also provides a bearing state prediction apparatus. The device comprises:

the first determining module is used for determining the service cycle of the bearing to be detected and determining the reference characteristic value of the bearing to be detected in the service cycle;

The second determining module is used for carrying out parameter updating on the initial characteristic change function according to the reference characteristic value to obtain a target characteristic change function;

the third determining module is used for predicting the predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function;

and the fourth determining module is used for predicting the state of the predicted characteristic value by adopting a target characteristic change function and a hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:

In a fourth aspect, the present application also provides a computer-readable storage medium. A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

In a fifth aspect, the present application also provides a computer program product. Computer program product comprising a computer program which, when executed by a processor, realizes the steps of:

According to the bearing state prediction method, the device, the computer equipment and the storage medium, the target characteristic change function is determined through the reference characteristic value of the bearing to be detected in the service period; predicting the characteristic value of the bearing to be detected through a target characteristic change function to obtain a predicted characteristic value of the bearing to be detected in a future time period; further, a target characteristic change function and a hidden Markov model are adopted to determine the predicted running state of the bearing to be detected in a future time period. Because the state of the bearing to be detected is predicted according to the combination of the target characteristic change function and the hidden Markov model when the predicted running state of the bearing to be detected in the future time period is determined in the process, compared with the process of predicting the state of the bearing only according to the hidden Markov model in the prior art, the running state of the bearing can be predicted more accurately. In addition, the target characteristic change function is obtained by updating parameters of the initial characteristic change function according to the reference characteristic value of the bearing to be detected, so that the reference characteristic value corresponding to the service period of the bearing to be detected is considered when the state of the bearing to be detected is predicted, and the accuracy of determining the predicted running state of the bearing to be detected in a future time period is further improved.

Drawings

Fig. 1 is an application environment diagram of a bearing state prediction method provided in an embodiment of the present application;

FIG. 2 is a flowchart of a method for predicting a bearing state according to an embodiment of the present disclosure;

FIG. 3 is a flowchart illustrating steps for determining a predicted operating state according to an embodiment of the present application;

FIG. 4 is a flowchart illustrating steps for determining a predicted operating state according to another embodiment of the present application;

FIG. 5 is a flowchart illustrating steps of a hidden Markov model training process according to an embodiment of the present application;

FIG. 6 is a flowchart of another method for predicting a bearing condition according to an embodiment of the present disclosure;

fig. 7 is a block diagram of a first bearing state predicting device according to an embodiment of the present application;

FIG. 8 is a block diagram of a second bearing condition predicting device according to an embodiment of the present disclosure;

FIG. 9 is a block diagram of a third bearing condition predicting device according to an embodiment of the present disclosure;

fig. 10 is a block diagram of a fourth bearing state predicting device according to an embodiment of the present disclosure;

FIG. 11 is a block diagram of a fifth bearing condition predicting device according to an embodiment of the present disclosure;

fig. 12 is an internal structural diagram of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.

It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application. In the description of the present application, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.

Based on the above situation, the bearing state prediction method provided in the embodiment of the present application may be applied to an application environment as shown in fig. 1. In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in FIG. 1. 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 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and data blocks. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The data block of the computer device is used for storing acquired data of the bearing state prediction method. 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 method of predicting a bearing condition.

The application discloses a bearing state prediction method, a device, computer equipment and a storage medium thereof, wherein parameter updating is carried out on an initial characteristic change function through a reference characteristic value of a bearing to be detected to obtain a target characteristic change function; predicting the characteristic value of the bearing to be detected according to the target characteristic change function to obtain a predicted characteristic value of the bearing to be detected in a future time period; further, a target characteristic change function and a hidden Markov model are adopted to conduct state prediction on the predicted characteristic value, and the predicted running state of the bearing to be detected in a future time period is obtained.

In an exemplary embodiment, as shown in fig. 2, fig. 2 is a flowchart of a method for predicting a bearing state according to an embodiment of the present application, and a method for predicting a bearing state is provided, and the method is applied to the computer device in fig. 1, for example, and is described as follows, including steps 201 to 204. Wherein:

step 201, determining a service cycle of the bearing to be detected, and determining a reference characteristic value of the bearing to be detected in the service cycle.

It should be noted that the use period refers to an nth time period in which the bearing to be detected is put into use, where a unit of the time period may be set and adjusted according to actual situations and requirements, and for example, the use period may refer to an nth month in which the bearing to be detected is put into use.

Further, when the reference characteristic value of the bearing to be detected in the service period needs to be determined, at least one bearing vibration amplitude of the bearing to be detected in the service period can be obtained, and then the reference characteristic value of the bearing to be detected in the service period is determined according to the bearing vibration amplitudes; the vibration amplitude of the bearing is used for representing the vibration amplitude of the bearing to be detected.

Specifically, since the bearing vibration amplitude of the bearing to be detected has uncertainty, that is, the bearing vibration amplitudes of the bearing to be detected corresponding to different service periods may be the same or different, when the reference characteristic value of the bearing to be detected in the service period needs to be determined, the following contents may be specifically included: collecting bearing vibration amplitude values of the bearing to be detected at each moment in the service period, and denoising the bearing vibration amplitude values to obtain denoised bearing vibration amplitude values; and further, determining a reference characteristic value of the bearing to be detected in the service period according to the vibration amplitude of the bearing after the denoising treatment.

Further, when the target bearing vibration amplitude after the denoising treatment needs to be obtained, the following specific contents may be included: the method comprises the steps of collecting the vibration amplitude of a bearing at each moment in the service period, obtaining at least one vibration amplitude of the bearing to be detected in the service period, denoising the vibration amplitude of each bearing by wavelet decomposition, and obtaining the vibration amplitude of each bearing after denoising.

Specifically, wavelet decomposition is carried out on each bearing vibration amplitude of the bearing to be detected to obtain wavelet coefficients corresponding to each bearing vibration amplitude, and the wavelet coefficients corresponding to each bearing vibration amplitude can be compared with a preset coefficient threshold value to obtain a size comparison result because the wavelet coefficients obtained by wavelet decomposition of the normal bearing vibration amplitude are larger and the wavelet coefficients obtained by wavelet decomposition of the noise amplitude are smaller; if the magnitude comparison result shows that the wavelet coefficient of the vibration amplitude of a certain bearing is larger than the coefficient threshold value, the vibration amplitude of the bearing corresponding to the wavelet coefficient belongs to the normal vibration amplitude of the bearing, so that the vibration amplitude of the bearing is reserved; if the magnitude comparison result shows that the wavelet coefficient of the vibration amplitude of a certain bearing is smaller than the coefficient threshold value, the vibration amplitude of the bearing corresponding to the wavelet coefficient belongs to the noise amplitude, so that the vibration amplitude of the bearing is cleared; and then, according to the size comparison result, denoising the bearing vibration amplitude at each moment in the service period to obtain the denoised bearing vibration amplitude.

Further, when the reference characteristic value of the bearing to be detected needs to be determined, the absolute average amplitude of the bearing vibration amplitude after denoising treatment can be extracted according to the change characteristics that the bearing vibration amplitude is firstly stable and then gradually rises, and the absolute average amplitude is used as the reference characteristic value of the bearing to be detected.

And 202, carrying out parameter updating on the initial characteristic change function according to the reference characteristic value to obtain a target characteristic change function.

The initial characteristic change function refers to a characteristic change function obtained by characteristic fitting of a historical characteristic value of the bearing to be detected in a historical period.

It should be noted that, in order to prevent the influence of the reference feature value of the use period on the predicted feature value from being unable to be considered when the predicted feature value of the bearing to be detected in the future time period is predicted by the initial feature change function of the bearing to be detected, the change trend of the bearing vibration amplitude of the bearing to be detected in the running process cannot be accurately reflected, so that the reference feature value of the bearing to be detected in the time to be detected needs to be updated for parameters of the initial feature change function to obtain the target feature change function.

Further, the method for jointly estimating the state parameters of the particle filter can update the parameters of the initial characteristic change function, and specifically can include the following steps: extracting a plurality of first characteristic values according to the prior probability of the reference characteristic value in the initial characteristic change function at the initial moment (namely, the moment when the initial characteristic change function starts to be updated), initializing the weight corresponding to each first characteristic value, extracting a plurality of second characteristic values according to an importance density function (taking the probability distribution of the prior probability as the importance density function) in the running process of the bearing, calculating to obtain the weight corresponding to each second characteristic value through a weight formula, carrying out normalization processing on the weight of each second characteristic value, determining effective second characteristic values according to the weight after normalization processing and a resampling calculation formula, and further carrying out parameter updating on the initial characteristic change function according to the effective second characteristic values.

In one embodiment of the present application, if the parameters of the initial feature change function are: when the initial feature change function needs to be updated, the feature change parameter a, the feature change parameter b, the feature change parameter c and the feature change parameter d need to be updated respectively, and the feature change parameter a, the feature change parameter b, the feature change parameter c and the feature change parameter d adopt the same updating process, so that the updating processes of the feature change parameter a, the feature change parameter b, the feature change parameter c and the feature change parameter d are not repeated one by one, and only the updating process of the feature change parameter a is developed to be described in detail:

for example, when the feature change parameter a of the initial feature change function needs to be updated, at time t=0 (the time when the initial feature change function starts to be updated), the prior probability p (x) of the reference feature value in the initial feature change function is calculated ₀ ) Extracting first characteristic values corresponding to N characteristic change parameters aAnd initializing the weight corresponding to each first characteristic value, namely +.>

Wherein p (x) ₀ ) Refers to the prior probability of the reference feature value in the initial feature change function, x ₀ Referring to the eigenvalue at time t=0, N refers to the number of samples extracted,refers to the ith each first characteristic value extracted at time t=0, ++>Refers to the weight corresponding to each first feature value of the ith extracted at the time of t=0.

Further, selecting prior probability distribution of reference feature values in initial feature change function in bearing operation processAs a function of the importance density->I.e. < -> Extracting second characteristic values corresponding to the N characteristic change parameters a according to the importance density function>

Wherein,refers to the prior probability distribution of the reference feature values in the initial feature variation function,refers to the importance density function, +.>Represents the ith second characteristic value extracted at time t,/->The i-th second characteristic value extracted at the time t-1 is represented by i, i represents the number of the second characteristic value, t represents the time t, z _1：t Representing the real characteristic value set from 1 to t.

Further, a weight value corresponding to each second characteristic value is obtained through calculation of a weight formula, and normalization processing is carried out on the weight value through a normalization formula, wherein the weight formula (1) is as follows:

wherein,weight corresponding to the ith second characteristic value extracted at time t is represented by +. >Representing the weight value, z corresponding to the ith second characteristic value extracted at the moment t-1 _t Representing the true characteristic value at time t +.>Represents the ith second characteristic value extracted at time t,/->Expressed in the known +.>Under the condition of (2) to obtain z _t Is a function of the probability of (1),

wherein, normalization formula (2) is as follows:

wherein,indicating the weight corresponding to the ith second characteristic value after normalization at the moment t, and ++>And the weight corresponding to the ith second characteristic value extracted at the moment t is represented.

Further, according to the resampling calculation formula (3), the number of the effective second characteristic values is obtained.

Wherein, the resampling calculation formula (3) is as follows:

wherein,representing the number of valid second eigenvalues, N representing the number of extracted second eigenvalues, +.>And the weight corresponding to the ith second characteristic value after normalization at the moment t is represented.

Further, comparing the number of the effective second characteristic values with a preset number threshold, and extracting the second characteristic values again by taking the normalized weight as an importance density function if the number of the effective second characteristic values is smaller than the preset number threshold; if the number of the effective second feature values is greater than or equal to the preset number threshold, updating the feature change parameter a of the initial feature change function according to the effective second feature values to obtain a value corresponding to the updated feature change parameter a, wherein the specific formula is as follows:

Wherein,representing the value corresponding to the updated characteristic change parameter a,/->Weight corresponding to the i-th valid second feature value extracted at time t,/->Representing the i-th valid second eigenvalue extracted at time t.

Further, the updated characteristic change parameter a is used for replacing the characteristic change parameter a in the initial characteristic change function, the value corresponding to the updated characteristic change parameter b is used for replacing the characteristic change parameter b in the initial characteristic change function, the value corresponding to the updated characteristic change parameter c is used for replacing the characteristic change parameter c in the initial characteristic change function, and the value corresponding to the updated characteristic change parameter d is used for replacing the characteristic change parameter d in the initial characteristic change function, so that the target characteristic change function is obtained.

For example, if the initial characteristic change function determined according to the historical characteristic value of the bearing to be detected in the historical period is:

wherein 1.767 is a characteristic change parameter d,0.308 is a characteristic change parameter a,0.254 is a characteristic change parameter b,2.673 is a characteristic change parameter c; therefore, in the state parameter joint estimation method according to the particle filtering, the initial feature change function is updated according to the reference feature value, so as to obtain an updated feature change parameter a, an updated feature change parameter b, an updated feature change parameter c and an updated feature change parameter d. And replacing the characteristic change parameter a, the characteristic change parameter b, the characteristic change parameter c and the characteristic change parameter d of the initial characteristic change function by using the updated characteristic change parameter a, the updated characteristic change parameter b, the updated characteristic change parameter c and the updated characteristic change parameter d to obtain the target characteristic change function.

And step 203, predicting a predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function.

The predicted characteristic value refers to a characteristic value of the bearing vibration amplitude of the bearing to be detected in at least one future period, which is predicted by the target characteristic change function.

It should be noted that, because the target feature change function is obtained by updating parameters of the initial feature change function through the reference feature value of the bearing to be detected, when the predicted feature value of the bearing to be detected in the future time period needs to be accurately obtained, the feature value of the bearing to be detected in the future time period can be predicted according to the target feature change function, so as to obtain the predicted feature value of the bearing to be detected in the future time period.

Further, when the characteristic value of the bearing to be detected in the future time period is predicted by using the target characteristic change function, if the predicted characteristic value of the bearing to be detected is greater than or equal to the failure threshold value, the prediction is stopped.

The failure threshold is a value preset according to the historical experience and actual condition of the staff.

In one embodiment of the present application, different time parameters are input into the target feature change function according to the target feature change function, so as to obtain an output result of the target feature change function, where the result is a predicted feature value of the bearing to be detected in different future time periods. Wherein the time parameter refers to at least one future period of the bearing to be detected after the period of use; for example, if the service period of the bearing to be detected is 96 months when the bearing to be detected is put into service, the 97 th month when the bearing to be detected is used, the 98 th month when the bearing to be detected is used, the 99 th month when the bearing to be detected is used, and the like are all future periods of the bearing to be detected.

For example, if the expression of the target feature change function is:

wherein t represents a time parameter, and F (t) represents a predicted characteristic value of the bearing to be detected in a t time period.

Further, if the value sequence of the given time parameter t is: {98,100,102,104,106,108,110,112,114,116,118,120,122,124,126,127,128,129,130,131}, substituting the value of t into the target characteristic change function to obtain a corresponding predicted characteristic value sequence as follows:

{2.92209636697330，3.00464706621310，3.1099735326498，3.24435960853055，

3.41582285040689，3.63459286152647，3.9137215963008，4.26986204799277，

4.72426177661193，5.30403055151030，6.04375773701856，6.98757591535332，

8.19179386381646，9.72825597133735，11.6886285192307，14.1898685483353，

17.3812015885365，21.4530245463378，22.6482649022715，33.2768739149030}

if the preset failure threshold is 23, the corresponding prediction characteristic value is 33.2768739149030 when t=131, and the prediction characteristic value of the bearing to be detected is 33.2768739149030>23, namely, the prediction characteristic value of the bearing to be detected is greater than the failure threshold, so that when the time parameter is 131, the prediction of the prediction characteristic value of the bearing to be detected in the future time period is stopped.

And 204, carrying out state prediction on the predicted characteristic value by adopting a target characteristic change function and a hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period.

It should be noted that, because the corresponding characteristic values of the bearing to be detected in different running states are different, the predicted running state of the bearing to be detected in the future time period can be determined according to the predicted characteristic value of the bearing to be detected in the future time period; specific: and obtaining a degradation trend function according to the target characteristic change function, further, updating a hidden Markov model through the degradation trend function to obtain a hidden Markov degradation model, and further, carrying out state prediction on the prediction characteristic value by using the hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

In an embodiment of the present application, when a predicted running state of a bearing to be detected in a future time period needs to be obtained, a hidden markov model may be updated according to a target feature change function to obtain a first hidden markov model, further, a state value of the bearing to be detected output by the first hidden markov model may be obtained by inputting a predicted feature value into the first hidden markov model, state value intervals corresponding to different running states are predetermined, and the running state corresponding to the bearing to be detected is determined according to the state value interval to which the state value belongs.

For example, if the state value interval corresponding to the first operation state is (0, 10) and the state value interval corresponding to the second operation state is (10, 20), and the state value of the bearing to be detected output by the first hidden markov model is known to be 12, the state value 12 belongs to the state value interval (10, 20), so that the operation state corresponding to the bearing to be detected is determined to be the second operation state.

In one embodiment of the present application, when a predicted running state of a bearing to be detected in a future time period needs to be obtained, a hidden markov model may be updated according to a target feature change function to obtain a second hidden markov model, and further, a state prediction result of the bearing to be detected output by the second hidden markov model may be obtained by inputting a predicted feature value into the second hidden markov model; the state prediction result records the probability that the bearing to be detected belongs to different operation states, and further, the operation state with the highest probability in the state prediction result is used as the operation state corresponding to the bearing to be detected.

For example, if the state prediction result of the bearing to be detected output by the second hidden markov model is [0.1,0.7,0.2], wherein 0.1 indicates that the running state of the bearing to be detected belongs to the first running state with a probability of 0.1,0.7 that the running state of the bearing to be detected belongs to the second running state with a probability of 0.7,0.2 indicates that the running state of the bearing to be detected belongs to the third running state with a probability of 0.2, and the second running state with the highest probability in the state prediction result is used as the running state corresponding to the bearing to be detected.

According to the bearing state prediction method, the target characteristic change function is determined through the reference characteristic value of the bearing to be detected in the service period; predicting the characteristic value of the bearing to be detected through a target characteristic change function to obtain a predicted characteristic value of the bearing to be detected in a future time period; further, a target characteristic change function and a hidden Markov model are adopted to determine the predicted running state of the bearing to be detected in a future time period. Because the state of the bearing to be detected is predicted according to the combination of the target characteristic change function and the hidden Markov model when the predicted running state of the bearing to be detected in the future time period is determined in the process, compared with the process of predicting the state of the bearing only according to the hidden Markov model in the prior art, the running state of the bearing can be predicted more accurately. In addition, the target characteristic change function is obtained by updating parameters of the initial characteristic change function according to the reference characteristic value of the bearing to be detected, so that the reference characteristic value corresponding to the use period of the bearing to be detected is considered when the state of the bearing to be detected is predicted, and the accuracy of determining the predicted running state of the bearing to be detected in a future time period is further improved.

In one embodiment, the prior art is not directed to the problem of accurately predicting the bearing operating condition for many emerging industries. To solve the above technical problem, the computer device of the present application may obtain, in a manner as shown in fig. 3, a predicted running state of a bearing to be detected in a future time period, including the following steps 301 to 303. Wherein:

step 301, determining a degradation trend function according to the target feature change function.

Wherein the degradation trend function refers to a function describing degradation of the bearing vibration amplitude of the bearing to be detected.

It should be noted that, because the vibration amplitude of the bearing to be detected is dynamically changed in the running process, and the vibration amplitude of the bearing to be detected is gradually increased along with the continuous increase of the service period of the bearing to be detected, that is, the bearing to be detected is gradually increased along with the continuous increase of the service period, and a certain degree of performance degradation is generated, so that the degradation trend function can be determined through the target feature change function for more accurately predicting the state of the predicted feature value.

In one embodiment of the present application, if the target feature change function is F (t _j ) Then the function F (t _j ) Analyzing the degradation condition of the bearing vibration amplitude of the bearing to be detected to obtain a degradation trend functionThe specific expression is as follows:

wherein,indicating the bearing to be detected at t _j Characteristic value of time>Representing a degradation threshold value of the bearing to be detected, +.>Representing a failure threshold value of the bearing to be detected and, < >>

And step 302, updating the hidden Markov model according to the degradation trend function to obtain a hidden Markov degradation model.

It should be noted that, in order to enable the hidden markov model to capture the degradation trend of the bearing to be detected more accurately, the hidden markov model may be updated according to the degradation trend function to obtain the hidden markov degradation model.

In one embodiment of the present application, the degradation trend function isUpdating the state probability transition matrix of the hidden Markov model to obtain an updated hidden Markov degradation model, wherein the state probability transition matrix of the hidden Markov degradation model is as follows:

wherein A 'is' _si (t _j ) Representing t _j The running state of the bearing to be detected at moment is in state S _i Time (in state S) _i The probability of (a) being the largest); t is t _su Indicating the moment corresponding to the transition of the running state of the bearing to be detected from u to u+1; a' _uu+1 (t _su ) Representing t _su Probability of transition of the operating state of the bearing to be detected from u to u+1 at the moment, and t _su ＜t _j ；a′ _ii+1 (t _j ) Representing t _j The probability of degradation transition of the bearing to be detected from state i to i+1 at a moment, which is equal to the output of the degradation trend function at this momentAnd epsilon _i The sum is: /> ε _i The deviation between the state transition probability and the output of the degradation trend function is represented as follows: />a _nn Representing the transition probability of the bearing to be detected being in a failure state, a _nn Has a value of 1, i.e. a _nn ＝1。

Further, the probability that the running state of the bearing to be detected is transferred from any state p to any state q is as follows: a' _pq (t) wherein 0.ltoreq.a' _pq (t) is less than or equal to 1, and

further described, due to A' _si (t _j ) Each element in (a) represents a state transition probability, the probability value is 1 at the maximum, thus A' _si (t _j ) The values of the elements of the system are not consistent with the detectionThe degradation tendency of the bearing grows unrestrained and stops increasing when the element value reaches 1.

And 303, carrying out state prediction on the predicted characteristic value by adopting a hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

It should be noted that, when it is required to obtain the predicted running state of the bearing to be detected in the future period of time, the following may be specifically included: and determining the initial running state of the bearing to be detected according to the predicted characteristic value of the bearing to be detected in the future time period, and further, carrying out state prediction on the predicted characteristic value according to the hidden Markov degradation model to obtain the candidate running state of the bearing to be detected in the future time period, and further, determining the candidate running state corresponding to the initial running state from the candidate running state, namely, the predicted running state of the bearing to be detected in the future time period.

According to the method for predicting the state of the bearing, the hidden Markov model is updated through the degradation trend function to obtain the hidden Markov degradation model, and the degradation trend function is obtained according to the characteristic change condition of the vibration amplitude of the bearing to be detected at the time to be detected, so that the hidden Markov degradation model obtained after the update of the degradation trend function can enable the hidden Markov degradation model to better predict the running state of the bearing to be detected in the future time period.

In an exemplary embodiment, when the state prediction is performed on the predicted feature value by using the hidden markov degradation model to obtain the predicted running state of the bearing to be detected in the future period of time, the following steps 401 and 402 may be included in the manner shown in fig. 4. Wherein:

step 401, determining an initial running state of the bearing to be detected in a future time period.

The initial running state refers to a running state of the bearing to be detected corresponding to a previous period of a future period. For example, if the future period is the 97 th month of bearing usage to be detected, the initial operating state refers to the 96 th month of bearing usage to be detected.

It should be noted that, when the initial running state of the future time period needs to be determined, if the future time period is the service period of the detection bearing, the initial running state of the bearing to be detected in the future time period refers to the running state of the bearing to be detected in the service period, and the initial running state of the bearing to be detected in the future time period can be determined according to the reference characteristic value corresponding to the service period of the bearing to be detected.

Further, there are many methods for determining the initial running state of the bearing to be detected in the future time period, for example, the feature state comparison table may be searched for data according to the feature state comparison table, and the initial running state corresponding to the reference feature value is determined, where the feature state comparison table records the initial running states corresponding to different reference feature values; or determining the initial running state corresponding to the reference characteristic value according to the hidden Markov model. It can be understood that the initial running state corresponding to the reference feature value can be determined in various manners, and the method for determining the initial running state corresponding to the reference feature value will not be described in detail herein, and the two methods will be described in detail below:

As an example, when the initial running state corresponding to the reference feature value is determined by searching the feature state lookup table according to the feature state lookup table, the following may be specifically included: and searching the data of the characteristic state comparison table according to the characteristic state comparison table, and finding an initial running state corresponding to the reference characteristic value, wherein the initial running state is the initial running state corresponding to the reference characteristic value.

As another example, when determining the initial operation state corresponding to the reference feature value according to the hidden markov model, the following may be specifically included: according to the observation value probability matrix of the hidden Markov model, a column corresponding to the reference characteristic value is found from the observation value probability matrix, and an initial running state corresponding to the reference characteristic value is determined according to the probability value in the column, wherein the initial running state is the initial running state corresponding to the reference characteristic value.

And step 402, carrying out state prediction on the predicted characteristic value according to the initial running state and the hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

When the predicted running state of the bearing to be detected in the future time period needs to be obtained, the predicted characteristic value can be input into the hidden Markov degradation model to obtain the candidate running state output by the hidden Markov degradation model; further, a predicted operating state corresponding to the initial operating state is determined from the candidate operating states.

In one embodiment of the present application, after the prediction feature value is input to the hidden markov degradation model, four candidate operation states output by the hidden markov degradation model are obtained as follows: s'. ₁ ，S′ ₂ ，S′ ₃ ，S′ ₄ The state transition probability matrixes corresponding to the four candidate operation states respectively are as follows:the state transition probability matrix corresponding to each of the four candidate operation states is specifically expressed as follows:

wherein,representing that the initial running state of the bearing to be detected is S ₁ Corresponding toState transition probability matrix of candidate operating states +.>Representing that the initial running state of the bearing to be detected is S ₂ State transition probability matrix of corresponding candidate operating states,/->Representing that the initial running state of the bearing to be detected is S ₃ A state transition probability matrix for the corresponding candidate operating state,representing that the initial running state of the bearing to be detected is S ₄ State transition probability matrix of corresponding candidate operating states,/->Indicating the bearing to be detected at t _j Characteristic value of time.

Further, when the predicted running state of the bearing to be detected needs to be determined, the method specifically comprises the following steps: the method for determining the predicted running state corresponding to the initial running state from the candidate running states specifically comprises the following steps: and finding a state transition probability matrix corresponding to the initial operation state from the state transition probability matrix of the candidate operation state, and further determining the predicted operation state of the bearing to be detected according to the element value in the state transition probability matrix corresponding to the initial operation state.

For example, when it is desired to determine the predicted running state of the bearing to be detected corresponding to the next future period corresponding to the usage period (the usage period may be represented by t, and the next future period may be represented by t+1, for example, if the usage period is 10 months, the next future period is 11 months.) the following may be specifically included: if the initial running state of the bearing to be detected is at S _i The probability of (2) is:wherein (1)>Indicating the bearing to be detected to be in the state S at the time t _i Probability of pi _i Indicating that the bearing to be detected is in the state S at the time t=1 _i Probability of b _i (o ₁ ) Indicating that at time t it is in state S _i The reference characteristic value of the lower corresponding bearing to be detected is o ₁ Is a probability of (2).

Further, according to a recurrence formula and a state transition probability matrix A 'corresponding to each candidate state' _si (t) obtaining the probability that the bearing to be detected belongs to each candidate state at the moment of t+1:

wherein,indicating that the bearing to be detected is positioned in each candidate state S 'at the time t+1' ₁ S′ ₂ ，，，S′ _n Probability of->Indicating that the bearing to be detected is positioned in each candidate state S 'at the moment t' ₁ S′ ₂ ，，，S′ _n Probability of A' _si (t) represents that the bearing to be detected is in the candidate state S 'at the moment t' _i Is a state transition probability matrix of (a).

Further, fromThe subscript of the maximum probability is the predicted running state of the bearing to be detected at the time t+1.

For example, when it is desired to determine the predicted running states of the bearing to be detected corresponding to the last two future periods of the usage period (the usage period may be represented by t, and the last two future periods may be represented by t+2, for example, if the usage period is 10 months, the next future period is 12 months.) the following may be specifically included: after the predicted characteristic value of the bearing to be detected at the time t+1 is input into the hidden Markov degradation model, the state transition probability corresponding to each candidate state of the bearing to be detected at the time t+1 is obtained through a recursion formula, and then the predicted running state of the bearing to be detected at the time t+1 is obtained from the state transition probability corresponding to each candidate state; further, after the predicted characteristic value of the bearing to be detected at the time t+2 is input into the hidden Markov degradation model, the state transition probability corresponding to each candidate state of the bearing to be detected at the time t+2 is obtained through a recurrence formula, and then the predicted running state of the bearing to be detected at the time t+2 is obtained from the state transition probability corresponding to each candidate state.

According to the bearing state prediction method, the initial running state of the bearing to be detected in the future time period is determined, a basis is provided for the subsequent state prediction by adopting the hidden Markov degradation model, and the obtained hidden Markov degradation model can accurately predict the predicted running state of the bearing to be detected in the future time period.

In an exemplary embodiment, the training process of the hidden Markov model may include the following steps 501 and 502 in the manner shown in FIG. 5. Wherein:

step 501, determining a historical characteristic value of a sample bearing in a historical period, and determining a historical running state corresponding to the historical characteristic value.

It should be noted that, in order to ensure that the hidden markov model can accurately predict the running state of the bearing to be detected, it is necessary to determine a historical characteristic value of the sample bearing in a historical period and determine a historical running state corresponding to the historical characteristic value, and use the historical characteristic value as training data of the hidden markov model to implement training of the hidden markov model.

Further, when the historical characteristic value of the sample bearing in the historical period needs to be determined, the bearing vibration amplitude of the bearing to be detected in the historical period can be collected, and the bearing vibration amplitude is subjected to denoising treatment by adopting wavelet decomposition, so that the denoised historical bearing vibration amplitude is obtained; and further, extracting the characteristics of the denoised historical bearing vibration amplitude, and determining the historical characteristic value of the sample bearing in the historical period.

Further, when the historical operation state corresponding to the historical characteristic value needs to be determined, a column corresponding to the predicted characteristic value can be found from the initial observation value probability matrix according to the initial observation value probability matrix of the initial model, and then the operation state corresponding to the row with the largest probability value in the column is found, wherein the operation state is the initial operation state of the bearing to be detected in a future time period.

Step 502, performing parameter adjustment on the state transition probability matrix parameters and the observation probability matrix parameters of the initial model according to the historical characteristic values and the historical running states corresponding to the historical characteristic values to obtain a hidden Markov model.

It should be noted that, in order to ensure that the hidden markov model can accurately predict the running state of the bearing to be detected, the historical running state of the bearing to be detected in the historical period and corresponding historical running state of the historical characteristic value are required to be used as training data of the initial model, so as to realize training of the initial model and obtain the trained hidden markov model.

Further illustratively, when a hidden Markov model is desired, the following may be included: the historical characteristic values and the historical running states corresponding to the historical characteristic values are input into an initial model, the initial model is updated and trained, and the parameters of the state transition probability matrix parameters and the observation value probability matrix parameters of the initial model are adjusted to obtain a trained hidden Markov model.

In one embodiment of the present application, if the historical feature value is: {1.59,1.59,1.63,1.68,1.73,1.85,2.30,2.524,2.524,3.090,3.446,3.859,4.432,4.903,5.558,6.321,7.210,8.245,9.452,10.859,12.918,15.987,20.112,25.652,33.096}, and the initial model is:

π＝[0.95,0.03,0.02,0.01]

Wherein A represents an initial state transition probability matrix, B represents an initial observation probability matrix, and pi represents an initial state probability vector.

Further, parameter adjustment is carried out on initial state transition probability matrix parameters and observation value probability matrix parameters of an initial model through historical characteristic values and historical running states of the bearing to be detected in a historical period, and an obtained hidden Markov model is:

wherein A 'represents a state transition probability matrix after parameter adjustment, and B' represents an observation value probability matrix after parameter adjustment.

According to the bearing state prediction method, the initial model is subjected to parameter adjustment through the historical characteristic values and the historical operation states corresponding to the historical characteristic values, so that training of the initial model is realized, the obtained hidden Markov model can accurately capture the operation states of the bearing to be detected, and a model foundation is provided for the subsequent state prediction of the bearing to be detected in a future time period.

In an exemplary embodiment, when it is required to obtain a predicted running state of the bearing to be detected in a future period of time, the following procedure may be specifically included, as shown in fig. 6:

step 601, determining a service cycle of the bearing to be detected, and determining a reference characteristic value of the bearing to be detected in the service cycle.

And step 602, carrying out parameter updating on the initial characteristic change function according to the reference characteristic value to obtain a target characteristic change function.

And 603, predicting a predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function.

Step 604, determining a degradation trend function according to the target feature change function.

And step 605, updating the hidden Markov model according to the degradation trend function to obtain a hidden Markov degradation model.

Step 606 determines an initial operating state of the bearing to be detected for a future period of time.

Step 607, inputting the predicted feature value into the hidden markov degradation model to obtain the candidate running state output by the hidden markov degradation model.

Step 608 determines a predicted operating state corresponding to the initial operating state from the candidate operating states.

In one embodiment of the present application, before the parameter updating of the initial feature change function, the construction process of the initial feature change function is further included, which specifically includes the following steps: determining a historical characteristic value of the sample bearing in a historical period; and determining characteristic change parameters corresponding to the bearing to be detected according to the historical characteristic values, and constructing an initial characteristic change function according to the characteristic change parameters.

It should be noted that, when it is required to determine the historical characteristic value of the sample bearing in the historical period, the following may be specifically included: collecting bearing vibration amplitude values of the bearing to be detected in a history period, denoising the initial bearing vibration amplitude values by adopting wavelet decomposition, and obtaining the denoised history bearing vibration amplitude values of the bearing to be detected in the history period; and further, extracting the characteristics of the vibration amplitude of the historical bearing, and determining the historical characteristic value of the sample bearing in the historical period.

Specifically, wavelet decomposition is performed on each bearing vibration amplitude of a bearing to be detected to obtain wavelet coefficients corresponding to each bearing vibration amplitude, and the wavelet coefficients obtained after wavelet decomposition of a normal bearing vibration amplitude are larger, and the wavelet coefficients obtained after wavelet decomposition of a noise amplitude are smaller, so that the wavelet coefficients corresponding to each bearing vibration amplitude can be compared with a preset coefficient threshold value to obtain a size comparison result for denoising the bearing vibration amplitude, and if the wavelet coefficient of a certain bearing vibration amplitude is larger than the coefficient threshold value, the bearing vibration amplitude corresponding to the wavelet coefficients belongs to the normal bearing vibration amplitude, so that the bearing vibration amplitude is reserved; if the wavelet coefficient of the vibration amplitude of a certain bearing is smaller than the coefficient threshold, the vibration amplitude of the bearing corresponding to the wavelet coefficient belongs to the noise amplitude, so that the vibration amplitude of the bearing is cleared; and then, according to the size comparison result, denoising the bearing vibration amplitude at each moment in the service period to obtain the denoised bearing vibration amplitude.

For example, if the vibration amplitude of the bearing to be detected in the history period is: { x ₁ ,x ₂ ,x ₃ ,…,x _N Further, extracting features of vibration amplitude of the historical bearing, and extracting absolute average amplitude of vibration amplitude of the historical bearingThe method comprises the following steps: />The absolute average amplitude +.>The historical characteristic value of the bearing to be detected in the historical period is obtained.

It should be noted that, when the feature change parameter corresponding to the bearing to be detected needs to be determined according to the historical feature value, and the initial feature change function is constructed according to the feature change parameter, the method specifically may include the following: according to the historical characteristic values, substituting the historical characteristic values into an initial characteristic change function with unknown parameters, carrying out parameter estimation through a maximum likelihood estimation formula, determining characteristic change parameters corresponding to the bearing to be detected, and further substituting the characteristic change parameters into the initial characteristic change function with unknown parameters to construct the initial characteristic change function.

In one embodiment of the present application, if the historical feature value is: {1.59,1.59,1.63,1.68,1.73,1.85,2.30,2.524,2.524,3.090,3.446,3.859,4.432,4.903,5.558,6.321,7.210,8.245,9.452,10.859,12.918,15.987,20.112,25.652,33.096}, the initial feature change function for which the parameters are unknown is:

Wherein the value of F (t) represents the characteristic value of the bearing to be detected at the moment t, a, b, c and d are characteristic change parameters, t represents time and t _threshold And (5) representing the performance degradation initial time of the bearing to be detected, and taking the value as the average value of the historical performance degradation time of the bearing to be detected.

Further, substituting the historical characteristic value into an initial characteristic change function with unknown parameters (namely, an initial characteristic change function without determining characteristic change parameters), estimating the characteristic change parameters through a maximum likelihood estimation formula, and determining the characteristic change parameters corresponding to the bearing to be detected, wherein the maximum likelihood estimation formula (8) is as follows:

where L (θ) represents a maximum likelihood function, θ may represent a feature variation parameter a, a feature variation parameter b, a feature variation parameter c, and feature variation parameters d, p (c) _i θ) means that under the parameter θ, the history feature value is x _i N represents the number of history eigenvalues, pi represents the running-in operation.

Further, the feature change parameters corresponding to the bearing to be detected are determined as follows: a=0.348, b=0.254, c=2.673, d= 1.767, substituting the feature variation parameters into an initial feature variation function with unknown parameters, and constructing the initial feature variation function, wherein the specific function expression is as follows:

Wherein the value of F (t) represents the characteristic value of the bearing to be detected at the moment t, and t represents time.

According to the bearing state prediction method, the target characteristic change function is determined through the reference characteristic value of the bearing to be detected in the service period; predicting the characteristic value of the bearing to be detected through a target characteristic change function to obtain a predicted characteristic value of the bearing to be detected in a future time period; further, a target characteristic change function and a hidden Markov model are adopted to determine the predicted running state of the bearing to be detected in a future time period. Because the state of the bearing to be detected is predicted according to the combination of the target characteristic change function and the hidden Markov model when the predicted running state of the bearing to be detected in the future time period is determined in the process, compared with the process of predicting the state of the bearing only according to the hidden Markov model in the prior art, the running state of the bearing can be predicted more accurately. In addition, the target characteristic change function is obtained by updating parameters of the initial characteristic change function according to the reference characteristic value of the bearing to be detected, so that the reference characteristic value corresponding to the service period of the bearing to be detected is considered when the state of the bearing to be detected is predicted, and the accuracy of determining the predicted running state of the bearing to be detected in a future time period is further improved.

It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a bearing state prediction device for realizing the above-mentioned bearing state prediction method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitations in the embodiments of the bearing state predicting device or devices provided below may be referred to the limitations of the bearing state predicting method hereinabove, and will not be repeated herein.

In one embodiment, as shown in fig. 7, there is provided a bearing state predicting apparatus including: the first determination module 10, the second determination module 20, the third determination module 30, and the fourth determination module 40, wherein:

the first determining module 10 is configured to determine a service cycle of the bearing to be detected, and determine a reference characteristic value of the bearing to be detected in the service cycle.

The second determining module 20 is configured to update parameters of the initial feature change function according to the reference feature value, so as to obtain a target feature change function.

And a third determining module 30, configured to predict a predicted feature value of the bearing to be detected in a future time period according to the target feature variation function.

And a fourth determining module 40, configured to perform state prediction on the predicted feature value by using the target feature variation function and the hidden markov model, so as to obtain a predicted running state of the bearing to be detected in a future time period.

According to the bearing state prediction device, the target characteristic change function is determined through the reference characteristic value of the bearing to be detected in the service period; predicting the characteristic value of the bearing to be detected through a target characteristic change function to obtain a predicted characteristic value of the bearing to be detected in a future time period; further, a target characteristic change function and a hidden Markov model are adopted to determine the predicted running state of the bearing to be detected in a future time period. Because the state of the bearing to be detected is predicted according to the combination of the target characteristic change function and the hidden Markov model when the predicted running state of the bearing to be detected in the future time period is determined in the process, compared with the process of predicting the state of the bearing only according to the hidden Markov model in the prior art, the running state of the bearing can be predicted more accurately. In addition, the target characteristic change function is obtained by updating parameters of the initial characteristic change function according to the reference characteristic value of the bearing to be detected, so that the reference characteristic value corresponding to the service period of the bearing to be detected is considered when the state of the bearing to be detected is predicted, and the accuracy of determining the predicted running state of the bearing to be detected in a future time period is further improved.

In one embodiment, as shown in fig. 8, there is provided a bearing state predicting apparatus in which the fourth determining module 40 includes: a first determination unit 41, a second determination unit 42, and a third determination unit 43, wherein:

a first determining unit 41 for determining a degradation trend function based on the target feature variation function.

A second determining unit 42 is configured to update the hidden markov model according to the degradation trend function, so as to obtain a hidden markov degradation model.

And a third determining unit 43, configured to perform state prediction on the predicted feature value by using a hidden markov degradation model, so as to obtain a predicted running state of the bearing to be detected in a future time period.

In one embodiment, as shown in fig. 9, there is provided a bearing state predicting apparatus in which the third determining unit 43 includes: a first determination subunit 431 and a second determination subunit 432, wherein:

a first determination subunit 431 is configured to determine an initial operation state of the bearing to be detected in a future period of time.

The second determining subunit 432 is configured to perform state prediction on the predicted feature value according to the initial running state and the hidden markov degradation model, so as to obtain a predicted running state of the bearing to be detected in a future time period.

The second determining subunit 432 is specifically configured to input the predicted feature value to the hidden markov degradation model, so as to obtain a candidate running state output by the hidden markov degradation model; a predicted operating state corresponding to the initial operating state is determined from the candidate operating states.

In one embodiment, as shown in fig. 10, there is provided a bearing state predicting apparatus, the bearing state predicting apparatus further comprising: a fifth determination module 50 and a training module 60, wherein:

and a fifth determining module 50, configured to determine a historical characteristic value of the sample bearing during the historical period, and determine a historical running state corresponding to the historical characteristic value.

The training module 60 is configured to perform parameter adjustment on the state transition probability matrix parameter and the observation probability matrix parameter of the initial model according to the historical feature value and the historical running state corresponding to the historical feature value, so as to obtain a hidden markov model.

In one embodiment, as shown in fig. 11, there is provided a bearing state predicting apparatus, the bearing state predicting apparatus further comprising: a sixth determination module 70 and a seventh determination module 80, wherein:

a sixth determining module 70 is configured to determine a historical characteristic value of the bearing to be detected in a historical period.

The seventh determining module 80 is configured to determine a feature variation parameter corresponding to the bearing to be detected according to the historical feature value, and construct an initial feature variation function according to the feature variation parameter.

The respective modules in the above-described bearing state predicting apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a terminal, and the internal structure thereof may be as shown in fig. 12. The computer device includes a processor, a memory, an input/output interface, a communication interface, a display unit, and an input means. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile 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 the operating system and computer programs in the non-volatile storage media. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a method of predicting a bearing condition. The display unit of the computer device is used for forming a visual picture, and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be a key, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.

It will be appreciated by those skilled in the art that the structure shown in fig. 12 is merely a block diagram of some of the structures associated with the present application and is not limiting of the computer device to which the present application may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

It should be noted that, the user information (including, but not limited to, user equipment information, user personal information, etc.) and the data (including, but not limited to, data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data are required to comply with the related laws and regulations and standards of the related countries and regions.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, data blocks, or other media used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The data blocks referred to in various embodiments provided herein may comprise at least one of relational data blocks and non-relational data blocks. The non-relational data blocks may include, but are not limited to, blockchain-based distributed data blocks, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples represent only a few embodiments of the present application, which are described in more detail and are not thereby to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.

Claims

1. A method of predicting a bearing condition, the method comprising:

determining the service cycle of a bearing to be detected, and determining a reference characteristic value of the bearing to be detected in the service cycle;

And carrying out state prediction on the predicted characteristic value by adopting the target characteristic change function and the hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period.

2. The method according to claim 1, wherein the performing state prediction on the predicted feature value by using the target feature change function and a hidden markov model to obtain a predicted running state of the bearing to be detected in a future time period includes:

and carrying out state prediction on the prediction characteristic value by adopting the hidden Markov degradation model to obtain the predicted running state of the bearing to be detected in a future time period.

3. The method according to claim 2, wherein said performing state prediction on the predicted feature value using the hidden markov degradation model to obtain a predicted running state of the bearing to be detected in a future time period includes:

4. A method according to claim 3, wherein said performing state prediction on said predicted eigenvalues based on said initial operating state and said hidden markov degradation model to obtain a predicted operating state of the bearing to be detected in a future time period comprises:

inputting the predicted characteristic value into the hidden Markov degradation model to obtain a candidate running state output by the hidden Markov degradation model;

and determining a predicted running state corresponding to the initial running state from the candidate running states.

5. The method of claim 1, wherein the training process of the hidden markov model comprises:

6. The method according to claim 1, wherein the method further comprises:

determining a historical characteristic value of the bearing to be detected in a historical period;

and determining the characteristic change parameters corresponding to the bearing to be detected according to the historical characteristic values, and constructing an initial characteristic change function according to the characteristic change parameters.

7. A bearing condition predicting apparatus, the apparatus comprising:

the third determining module is used for predicting a predicted characteristic value of the bearing to be detected in a future time period according to the target characteristic change function;

and the fourth determining module is used for predicting the state of the predicted characteristic value by adopting the target characteristic change function and the hidden Markov model to obtain the predicted running state of the bearing to be detected in a future time period.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.

9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.