CN113158346A - Preventive maintenance method based on cloud computing - Google Patents

Preventive maintenance method based on cloud computing Download PDF

Info

Publication number
CN113158346A
CN113158346A CN202110530260.XA CN202110530260A CN113158346A CN 113158346 A CN113158346 A CN 113158346A CN 202110530260 A CN202110530260 A CN 202110530260A CN 113158346 A CN113158346 A CN 113158346A
Authority
CN
China
Prior art keywords
feature vector
performance
current
vector
environment
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110530260.XA
Other languages
Chinese (zh)
Other versions
CN113158346B (en
Inventor
刘积英
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SHANDONG ZHONGZHI ELECTRONICS CO Ltd
Original Assignee
Hangzhou Jiying Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangzhou Jiying Technology Co ltd filed Critical Hangzhou Jiying Technology Co ltd
Priority to CN202110530260.XA priority Critical patent/CN113158346B/en
Publication of CN113158346A publication Critical patent/CN113158346A/en
Application granted granted Critical
Publication of CN113158346B publication Critical patent/CN113158346B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/80—Management or planning

Abstract

The application relates to the field of cloud computing, in particular to a preventive maintenance method based on cloud computing. The method is based on cloud computing, a deep neural network model is used for predicting the sealing performance of the sealing ring from environmental information of the sealing ring in the using process, and then whether preventive maintenance is needed to be carried out on industrial equipment with the sealing ring is analyzed, wherein the information of the sealing ring in the using process comprises but is not limited to humidity, temperature, cooling water quality, altitude and the like. Therefore, the preventive maintenance prediction is carried out based on the cloud computing and the artificial intelligence technology, so that the effectiveness of the preventive maintenance is effectively improved, the stability of the operation of the industrial equipment is ensured, the frequency and the times of the preventive maintenance can be reduced, and unnecessary preventive maintenance is reduced.

Description

Preventive maintenance method based on cloud computing
Technical Field
The present application relates to the field of smart cloud computing, and more particularly, to a preventive maintenance method based on cloud computing, a preventive maintenance system based on cloud computing, and an electronic device.
Background
In the cloud computing, a huge data computing processing program is decomposed into countless small programs through a network cloud, and then the small programs are processed and analyzed through a system consisting of a plurality of servers to obtain results and the results are returned to a user. The cloud computing is a result of hybrid evolution and leap of computer technologies such as distributed computing, utility computing, parallel computing, network storage, hot backup redundancy and virtualization and the like, and the core idea is to uniformly manage a large number of computers for network connection, so that a computing resource platform is constructed to provide demand services for users.
In recent years, 10% of all cancelled or delayed flights are due to unscheduled maintenance events, which results in over 80 billion dollars of cost expenditure for the worldwide airline industry each year. However, the performance and reliability of the existing preventive maintenance system are not good due to the influence of many factors and unpredictable bottlenecks involved in the special maintenance of the aircraft, such as humidity, air temperature, cooling water quality and altitude.
For example, in aircraft equipment, if a seal ring fails or contaminants enter the oil filter system, the oil filter will fail before the specified limit is reached, thereby damaging the hydraulic system and other transmission components. However, the performance of the sealing ring is affected by various factors such as humidity, air temperature, cooling water quality and altitude, and it is difficult for the existing preventive maintenance system to make a judgment of when the preventive maintenance is accurately performed and when the preventive maintenance is required.
It is therefore desirable to be able to provide a new preventive maintenance solution.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. Embodiments of the present application provide a preventive maintenance method, system and electronic device based on cloud computing, which uses a deep neural network model to predict sealing performance of a sealing ring from environmental information during the use of the sealing ring based on cloud computing, and further analyzes whether an industrial device having the sealing ring needs preventive maintenance, wherein the information during the use of the sealing ring includes, but is not limited to, humidity, temperature, cooling water quality, altitude and the like. Therefore, the preventive maintenance prediction is carried out based on the cloud computing and the artificial intelligence technology, so that the effectiveness of the preventive maintenance is effectively improved, the stability of the operation of the industrial equipment is ensured, the frequency and the times of the preventive maintenance can be reduced, and unnecessary preventive maintenance is reduced.
According to an aspect of the present application, there is provided a cloud computing-based preventive maintenance method, including:
acquiring sealing ring performance data of a last routine maintenance test of an airplane at a cloud server for cloud computing and acquiring environmental data of the airplane in the flying process from a data recording device positioned at the end side of the airplane;
passing the seal ring performance data through a hidden Markov model to obtain a performance input vector;
passing the environment data through a hidden Markov model to obtain an environment input vector;
respectively passing the performance input vector and the environment input vector through a deep neural network to obtain a current performance characteristic vector and a current environment characteristic vector;
obtaining a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server;
calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative of the feature values of the current performance feature vector and the plurality of historical performance feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current performance feature vector by the sum of all the current performance feature vectors and the weighted sum values of the plurality of historical performance feature vectors;
calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient, wherein the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative number of the feature value of each position of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current environment feature vector by the sum of all the weighted sum values of the current environment feature vector and the plurality of historical environment feature vectors;
weighting the feature values of the positions of the current performance feature vector by the first weighting coefficient to obtain a first weighted vector, and weighting the feature values of the positions of the current environment feature vector by the second weighting coefficient to obtain a second weighted feature vector;
concatenating the first weighted feature vector and the second weighted feature vector to obtain a classified feature vector; and
and passing the classified feature vectors through a classifier to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
According to another aspect of the present application, there is provided a cloud computing-based preventative maintenance system, including:
the device comprises an environmental data acquisition unit, a data recording unit and a data processing unit, wherein the environmental data acquisition unit is used for acquiring sealing ring performance data of the last routine maintenance test of the airplane at a cloud server based on cloud computing and acquiring environmental data of the airplane in the flying process from a data recording device positioned at the end side of the airplane;
the performance input vector generating unit is used for enabling the sealing ring performance data to pass through a hidden Markov model so as to obtain a performance input vector;
the environment input vector generating unit is used for enabling the environment data obtained by the environment data obtaining unit to pass through a hidden Markov model so as to obtain an environment input vector;
the performance environment feature vector generation unit is used for enabling the performance input vector and the environment input vector to pass through a deep neural network respectively so as to obtain a current performance feature vector and a current environment feature vector;
a historical performance feature vector generation unit, configured to obtain a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server;
a first weighting coefficient generating unit, configured to calculate a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient, where the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values raised by negatives of feature values of respective positions of the current performance feature vector and the plurality of historical performance feature vectors to obtain a plurality of weighted sum values, and then divide the weighted sum value of the current performance feature vector by a sum of weighted sum values of all the current performance feature vectors and the plurality of historical performance feature vectors;
a second weighting factor generation unit configured to calculate a global Softmax-like function value of the current environment feature vector to obtain a second weighting factor, where the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values raised by negatives of feature values of respective positions of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and then divide the weighted sum value of the current environment feature vector by a sum of weighted sum values of all the current environment feature vectors and the plurality of historical environment feature vectors;
a weighted feature vector generation unit, configured to weight feature values of respective positions of the current performance feature vector by the first weighting coefficient to obtain a first weighted vector, and weight feature values of respective positions of the current environment feature vector by the second weighting coefficient to obtain a second weighted feature vector;
a classification feature vector generation unit configured to cascade the first weighted feature vector and the second weighted feature vector to obtain a classification feature vector; and
and the classification result generating unit is used for enabling the classification characteristic vectors to pass through a classifier so as to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
According to still another aspect of the present application, there is provided an electronic apparatus including: a processor; and a memory having stored therein computer program instructions that, when executed by the processor, cause the processor to perform the cloud computing-based preventative maintenance method as described above.
Compared with the prior art, the embodiment of the application provides a preventive maintenance method, a system and an electronic device based on cloud computing, which use a deep neural network model to predict the sealing performance of a sealing ring from environmental information in the use process of the sealing ring based on cloud computing, and further analyze whether industrial equipment with the sealing ring needs preventive maintenance or not, wherein the information in the use process of the sealing ring includes but is not limited to humidity, temperature, cooling water quality, altitude and the like. Therefore, the preventive maintenance prediction is carried out based on the cloud computing and the artificial intelligence technology, so that the effectiveness of the preventive maintenance is effectively improved, the stability of the operation of the industrial equipment is ensured, the frequency and the times of the preventive maintenance can be reduced, and unnecessary preventive maintenance is reduced.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.
Fig. 1 is an application scenario diagram of a cloud computing-based preventative maintenance method according to an embodiment of the present application.
Fig. 2 is a flowchart of the cloud computing-based preventative maintenance method according to an embodiment of the present application.
Fig. 3 is a schematic architecture diagram of the cloud computing-based preventative maintenance method according to an embodiment of the present application.
Fig. 4 is a flowchart of passing the seal ring performance data through a hidden markov model to obtain a performance input vector in the cloud computing-based preventative maintenance method according to the embodiment of the present application.
Fig. 5 is a flowchart of passing the environment data through a hidden markov model to obtain an environment input vector in the cloud computing-based preventative maintenance method according to the embodiment of the present application.
Fig. 6 is a flowchart of passing the classification feature vector through a classifier to obtain a classification result in the cloud computing-based preventative maintenance method according to the embodiment of the present application.
Fig. 7 is a block diagram of a cloud-computing-based preventative maintenance system according to an embodiment of the present application.
Fig. 8 is a block diagram of a performance input vector generation unit in a cloud computing-based preventative maintenance system according to an embodiment of the present application.
Fig. 9 is a block diagram of an environment input vector generation unit in a cloud computing-based preventative maintenance system according to an embodiment of the present application.
Fig. 10 is a block diagram of a classification result generation unit in the cloud computing-based preventative maintenance system according to an embodiment of the present application.
FIG. 11 is a block diagram of an electronic device according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Overview of a scene
As described above, since the sealing effect of the hydraulic circuit oil filter may be affected by factors such as air humidity, temperature, cooling water quality, and altitude, it is desirable to provide a preventive maintenance method for predicting a possible failure based on statistically preset information before a theoretical failure point when the number of flight times exceeds a certain threshold.
Based on this, the applicant of the present application considers that the preventive maintenance of such a fault is mainly to build a statistical model according to the existing information, so as to fully mine the statistical features in the existing information, and finally to perform normalization on the basis of the mined features to obtain a deterministic conclusion, so that the method is very suitable for using the currently popular deep neural network model based on the statistical model.
Therefore, the applicant of the present application considers using a deep neural network model to predict the sealing performance of the sealing ring from environmental information during the use of the sealing ring, i.e., humidity, temperature, cooling water quality, and altitude. Here, in order to simplify the consideration of the model, the applicant of the present application adopts a manner of prediction for each flight segment, that is, in routine maintenance of the aircraft before each flight segment, firstly, the service performance of the current sealing ring is acquired, then environmental information, i.e., humidity, temperature, cooling water quality and altitude, is recorded during the flight, and prediction is performed based on the information in the next routine maintenance after the aircraft lands.
Based on this, in the technical scheme of this application, at first, the sealing washer performance data that the high in the clouds server was tested when obtaining aircraft routine maintenance last time, for example can be the data that the maintainer uploaded to the high in the clouds server to from the environmental data of the data recorder record in the terminal side of aircraft in-process of flight, and convert these two data into high-dimensional vector through hidden markov's model respectively. Here, the hidden markov model is used because the distance between vectors can be considered in the vectors into which the model is converted, and whether the seal ring performance data obtained by the test or the environmental data has a certain relationship with the geographical position, so that the distance between the high-dimensional vectors obtained by each conversion of the seal ring performance data corresponds to the distance between the geographical positions where the data is obtained, and the distance between the high-dimensional vectors obtained by each conversion of the environmental data corresponds to the voyage distance, so that the model can contain as much information about the voyage as possible before feature extraction.
Then, the performance input vector and the environment input vector obtained by conversion are respectively passed through a deep neural network to obtain a performance feature vector and an environment feature vector, where it is desirable to be able to set appropriate weights thereof before fusing the performance feature vector and the environment feature vector. And because the characteristics corresponding to the two vectors are substantially changed along with the whole flight process, namely, are related to data in the life cycle of the airplane when in use, the technical scheme of the application further utilizes a characteristic memory mechanism to acquire the performance characteristic vector and the environmental characteristic in the cloud serverHistorical data of the vectors, namely a plurality of historical performance characteristic vectors and a plurality of historical environment characteristic vectors, and calculating a softmax-like function, namely sigma, of the current performance characteristic vector and the current environment characteristic vector relative to the historical performance characteristic vector and the historical environment characteristic vector respectivelyiexp(-xi)/∑j∑iexp (-xi), that is, each performance feature vector is subjected to exp (-xi) of each position and summed, and then the sum of all current and historical performance feature vectors is divided by the number of the current performance feature vector, so as to obtain the weighting coefficient of the performance feature vector, wherein the calculation mode of the environment feature vectors is the same. This is equivalent to calculating the weight of the probability distribution of the current information under the global angle, for example, if the temperature and humidity corresponding to the current flight has a large change relative to the global, the weight of the temperature and humidity should be increased inevitably. It should be noted that, here, the current performance feature vector and the current environment feature vector are also stored in the cloud server after being obtained.
Thus, after the obtained weighting coefficient is multiplied by the value of each position of the current performance characteristic vector and the current environment characteristic vector, the weighted characteristic vectors are cascaded to obtain a classified characteristic vector, and a classification result is obtained through a classifier, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring, namely whether a fault is possible or not.
Based on this, the present application proposes a preventive maintenance method based on cloud computing, which includes: acquiring sealing ring performance data of a last routine maintenance test of an airplane at a cloud server for cloud computing and acquiring environmental data of the airplane in the flying process from a data recording device positioned at the end side of the airplane; passing the seal ring performance data through a hidden Markov model to obtain a performance input vector; passing the environment data through a hidden Markov model to obtain an environment input vector; respectively passing the performance input vector and the environment input vector through a deep neural network to obtain a current performance characteristic vector and a current environment characteristic vector; obtaining a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server; calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative of the feature values of the current performance feature vector and the plurality of historical performance feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current performance feature vector by the sum of all the current performance feature vectors and the weighted sum values of the plurality of historical performance feature vectors; calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient, wherein the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative number of the feature value of each position of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current environment feature vector by the sum of all the weighted sum values of the current environment feature vector and the plurality of historical environment feature vectors; weighting the feature values of the positions of the current performance feature vector by the first weighting coefficient to obtain a first weighted vector, and weighting the feature values of the positions of the current environment feature vector by the second weighting coefficient to obtain a second weighted feature vector; concatenating the first weighted feature vector and the second weighted feature vector to obtain a classified feature vector; and enabling the classification feature vectors to pass through a classifier to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
Fig. 1 is an application scenario diagram of a cloud computing-based preventative maintenance method according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, seal ring performance data of a last routine maintenance test of an aircraft (e.g., V as illustrated in fig. 1) is acquired at a cloud server (C) for cloud computing, and environmental data of the aircraft during flight is acquired from a data recording device located at an end side of the aircraft; then, the various pieces of environmental data and the performance data of the sealing ring are input into the cloud server deployed with a preventive maintenance algorithm based on cloud computing, wherein the cloud server can process the performance data of the sealing ring and the environmental data based on the preventive maintenance algorithm based on cloud computing to generate a prediction result, and the prediction result represents a prediction result of the sealing effect of the current sealing ring, namely whether a fault is likely to occur. In turn, a determination may be made whether preventative maintenance is needed based on the prediction.
Although, in this application scenario, the seal ring is taken as an example of a seal ring in a hydraulic circuit oil filter on an aircraft, it should be understood by those skilled in the art that an analysis object of the cloud computing-based preventive maintenance method according to the present application is not limited to the seal ring in the hydraulic circuit oil filter on the aircraft, but is also applicable to other seal rings in a hydraulic circuit oil filter for a vehicle or other seal rings affected by the environment, and is not limited to this application.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary method
Fig. 2 is a flowchart of the cloud computing-based preventative maintenance method according to an embodiment of the present application. As shown in fig. 2, the cloud computing-based preventative maintenance method according to the embodiment of the present application includes the steps of: s110, acquiring sealing ring performance data of the last routine maintenance test of the airplane and acquiring environment data of the airplane in the flying process from a data recording device positioned at the end side of the airplane at a cloud server for cloud computing; s120, passing the performance data of the sealing ring through a hidden Markov model to obtain a performance input vector; s130, passing the environment data through a hidden Markov model to obtain an environment input vector; s140, respectively passing the performance input vector and the environment input vector through a deep neural network to obtain a current performance characteristic vector and a current environment characteristic vector; s150, acquiring a plurality of historical performance characteristic vectors and a plurality of historical environment characteristic vectors from the cloud server; s160, calculating a global Softmax-like function value of the current performance characteristic vector to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance characteristic vector is equal to a weighted sum of natural constant exponential function values which are respectively powered by the negative numbers of characteristic values of all positions of the current performance characteristic vector and the plurality of historical performance characteristic vectors to obtain a plurality of weighted sum values, and then dividing the weighted sum value of the current performance characteristic vector by the sum value of all the weighted sum values of the current performance characteristic vector and the plurality of historical performance characteristic vectors; s170, calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient, wherein the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values which are respectively powered by the negative numbers of feature values of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and then dividing the weighted sum value of the current environment feature vector by the sum value of all the weighted sum values of the current environment feature vector and the plurality of historical environment feature vectors; s180, weighting the feature values of the positions of the current performance feature vector by the first weighting coefficient to obtain a first weighted vector, and weighting the feature values of the positions of the current environment feature vector by the second weighting coefficient to obtain a second weighted feature vector; s190, cascading the first weighted feature vector and the second weighted feature vector to obtain a classified feature vector; and S200, enabling the classification characteristic vectors to pass through a classifier to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
Fig. 3 illustrates an architecture diagram of the cloud computing-based preventative maintenance method according to an embodiment of the present application. As shown IN fig. 3, IN this network architecture, first, seal ring performance data of a last routine maintenance test of an aircraft (for example, IN1 as illustrated IN fig. 3) is acquired at a cloud server for cloud computing and environmental data of the aircraft during flight (for example, IN2 as illustrated IN fig. 3) is acquired from a data recording device located at an end side of the aircraft; next, passing the seal ring performance data through a hidden markov model (e.g., an HMM as illustrated in fig. 3) to obtain a performance input vector (e.g., V1 as illustrated in fig. 3); then, passing the environment data through a hidden markov model to obtain an environment input vector (e.g., V2 as illustrated in fig. 3); then, passing the performance input vector and the environment input vector through a deep neural network (e.g., DNN as illustrated in fig. 3) to obtain a current performance feature vector (e.g., V3 as illustrated in fig. 3) and a current environment feature vector (e.g., V4 as illustrated in fig. 3), respectively; then, obtaining a plurality of historical performance feature vectors (e.g., V5 as illustrated in fig. 3) and a plurality of historical environment feature vectors (e.g., V6 as illustrated in fig. 3) from the cloud server; then, calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient (e.g., as illustrated in fig. 3, C1), wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values raised by the negative of the feature values of the current performance feature vector and the plurality of historical performance feature vectors, respectively, to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current performance feature vector by the sum of the weighted sum values of all the current performance feature vectors and the plurality of historical performance feature vectors; then, calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient (e.g., as illustrated in fig. 3, C2), where the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values raised by the negative numbers of the feature values of the respective positions of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current environment feature vector by the sum of the weighted sum values of all the current environment feature vectors and the plurality of historical environment feature vectors; then, weighting the feature values of the respective positions of the current performance feature vector by the first weighting coefficient to obtain a first weighting vector (e.g., V7 as illustrated in fig. 3); then, weighting the feature values of the respective positions of the current environment feature vector by the second weighting coefficients to obtain a second weighted feature vector (e.g., as illustrated in fig. 3 as V8); then, concatenating the first weighted feature vector and the second weighted feature vector to obtain a classified feature vector (e.g., V as illustrated in fig. 3); the classified feature vectors are then passed through a classifier to obtain a classification result (e.g., K as illustrated in fig. 3), wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
In step S110, seal ring performance data of a last routine maintenance test of an aircraft and environmental data of the aircraft during flight are acquired from a data recording device located on an end side of the aircraft at a cloud server for cloud computing. As mentioned above, the sealing effect of the oil filter due to the hydraulic circuit may be affected by factors such as air humidity, temperature, cooling water quality, and altitude. Therefore, in the embodiment of the present application, in order to determine the sealing effect of the current sealing ring, various items of data of the aircraft are acquired, which include: the method comprises the steps of obtaining sealing ring performance data of a routine maintenance test on an airplane and obtaining environment data of the airplane during the flight process from a data recording device located at the end side of the airplane. Accordingly, in the embodiments of the present application, the environmental data includes, but is not limited to, humidity, temperature, cooling water quality, altitude, and the like.
In step S120, the seal ring performance data is passed through a hidden markov model to obtain a performance input vector. That is, the hidden markov model is used to convert the environmental data and the sealing ring performance data into high-dimensional vectors, and here, the hidden markov model is used because the distance between the vectors can be considered in the vectors converted by the model, and no matter the sealing ring performance data obtained by the test or the environmental data has a certain relation with the geographical position, so that the distance between the high-dimensional vectors obtained by each conversion of the sealing ring performance data corresponds to the distance between the geographical positions where the data is obtained, and the distance between the high-dimensional vectors obtained by each conversion of the environmental data corresponds to the voyage distance, so that the high-dimensional information about the voyage can be included as much as possible before the feature extraction is performed.
Accordingly, in the embodiment of the present application, the process of passing the sealing ring performance data through the hidden markov model to obtain the performance input vector includes: firstly, acquiring geographic position data of the airplane in the flight process; then, based on the corresponding relationship between the distance between the high-dimensional vectors obtained by the conversion of the sealing ring performance data and the distance between the high-dimensional vectors obtained by the conversion of the geographical position data, the sealing ring performance data is passed through a hidden markov model to obtain the performance input vector.
Fig. 4 illustrates a flow chart of passing the seal ring performance data through a hidden markov model to obtain a performance input vector in a cloud computing-based preventative maintenance method according to an embodiment of the present application. As shown in fig. 4, passing the seal ring performance data through a hidden markov model to obtain a performance input vector, includes: s121, acquiring geographic position data of the airplane in the flight process; and S122, based on the corresponding relation between the distance between the high-dimensional vectors obtained by converting the sealing ring performance data and the distance between the high-dimensional vectors obtained by converting the geographical position data, the sealing ring performance data is processed by a hidden Markov model to obtain the performance input vector.
In step S130, the environment data is passed through a hidden markov model to obtain an environment input vector. That is, the environment input vector is obtained from the environment data by a hidden markov model. Accordingly, in an embodiment of the application, the process of passing the environment data through a hidden markov model to obtain an environment input vector comprises: s151, acquiring range distance data of the airplane in the flying process; and S152, based on the corresponding relation between the distance between the high-dimensional vectors obtained by the environmental data conversion and the distance between the high-dimensional vectors obtained by the voyage distance data conversion, the environmental data is processed by a hidden Markov model to obtain the environmental input vector, as shown in FIG. 5.
In step S140, the performance input vector and the environment input vector are respectively passed through a deep neural network to obtain a current performance feature vector and a current environment feature vector. That is, the performance input vector and the environment input vector are processed with a deep neural network to extract high-dimensional implicit features in the performance input vector and the environment input vector.
In one particular example of the present application, the deep neural network may be implemented as a deep fully-connected network, i.e., the deep neural network is composed of a plurality of fully-connected layers that are capable of sufficiently mining and utilizing information of various locations in the performance input vector and the environmental input vector to generate the current performance feature vector and the current environmental feature vector.
In step S150, a plurality of historical performance feature vectors and a plurality of historical environment feature vectors are obtained from the cloud server. When the seal ring performance characteristics and the environmental characteristics are fused for preventive analysis and judgment, it is desirable to set an appropriate weight thereof. Since the corresponding features of the two vectors substantially vary along the whole flight process, that is, are related to the data of the airplane in the life cycle when the airplane is in use, the technical solution of the present application further utilizes the mechanism of feature memory.
Specifically, a plurality of historical performance feature vectors and a plurality of historical environment feature vectors are first obtained from the cloud server, where the historical performance feature vectors and the historical environment feature vectors may be obtained based on the data processing process illustrated in step S110 to step S140, and details thereof are not repeated.
In step S160, a global Softmax-like function value of the current performance feature vector is calculated to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values raised by the negative of the feature values of the current performance feature vector and the plurality of historical performance feature vectors, respectively, to obtain a plurality of weighted sum values, and the weighted sum value of the current performance feature vector is divided by the sum of the weighted sum values of all the current performance feature vectors and the plurality of historical performance feature vectors.
Specifically, in the embodiment of the present application, the process of calculating the global class Softmax function value of the current performance feature vector to obtain the first weighting coefficient includes: calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient according to the following formula: a1 ═ Σiexp(-xi)/∑j∑iexp (-xi), a1 represents the first weighting coefficient, xi represents the eigenvalue of each position in the current performance eigenvector, and j represents each position of the historical performance eigenvector.
In step S170, a global Softmax-like function value of the current environment feature vector is calculated to obtain a second weighting coefficient, where the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponent function values raised to the negative of the feature values of the current environment feature vector and the plurality of historical environment feature vectors, respectively, to obtain a plurality of weighted sum values, and the weighted sum value of the current environment feature vector is divided by a sum of weighted sum values of all the current environment feature vectors and the plurality of historical environment feature vectors. That is, the global Softmax-like function value of the current environment feature vector is calculated to obtain the second weighting coefficient. That is, the global Softmax-like function value of the current environment feature vector is calculated to obtain the second weighting coefficient.
Specifically, in this embodiment of the present application, the process of calculating the global class Softmax function value of the current environment feature vector to obtain the second weighting coefficient includes: calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient according to the following formula: a2 ═ Σiexp(-xi)/∑j∑iexp (-xi), a2 represents the second weighting coefficient, xi represents the feature value of each position in the current environment feature vector, and j represents each position in the historical environment feature vector.
That is, in the technical scheme of the present application, a mechanism of feature memory is further utilized to obtain the feature information from the cloud serverTaking historical data of the performance characteristic vector and the environment characteristic vector, namely a plurality of historical performance characteristic vectors and a plurality of historical environment characteristic vectors, and respectively calculating a similar Softmax function of the current performance characteristic vector and the current environment characteristic vector relative to the same, namely sigmaiexp(-xi)/∑j∑iexp (-xi), that is, each performance feature vector is subjected to exp (-xi) of each position and summed, and then the sum of all current and historical performance feature vectors is divided by the number of the current performance feature vector, so as to obtain the weighting coefficient of the performance feature vector, wherein the calculation mode of the environment feature vectors is the same. This is equivalent to calculating the weight of the probability distribution of the current information under the global angle, for example, if the temperature and humidity corresponding to the current flight has a large change relative to the global, the weight of the temperature and humidity should be increased inevitably. It should be noted that, here, the current performance feature vector and the current environment feature vector are also stored in the cloud server after being obtained.
In step S180, the feature values of the positions of the current performance feature vector are weighted by the first weighting factor to obtain a first weighted vector, and the feature values of the positions of the current environment feature vector are weighted by the second weighting factor to obtain a second weighted feature vector. That is, the feature values of the respective positions of the current performance feature vector are weighted by the first weighting coefficient to obtain a first weighted vector, and similarly, the feature values of the respective positions of the current environmental feature vector are weighted by the second weighting coefficient to obtain a second weighted feature vector.
In step S190, the first weighted feature vector and the second weighted feature vector are concatenated to obtain a classified feature vector.
In step S200, the classified feature vectors are passed through a classifier to obtain a classification result, where the classification result represents a prediction result of the sealing effect of the current sealing ring. That is, after the classification feature vector is obtained, the classification feature vector is passed through a classifier to obtain a classification result, and the classification result represents a prediction result of the sealing effect of the current sealing ring, that is, whether a failure is likely to occur.
Specifically, in the embodiment of the present application, the process of passing the classification feature vector through a classifier to obtain a classification result includes: s201, enabling the classification characteristic vectors to pass through a Softmax function to obtain a first probability that the current sealing ring fails and a second probability that the current sealing ring cannot fail; and S202, determining the classification result based on the first probability and the second probability, as shown in fig. 6.
In summary, a cloud computing-based preventive maintenance method based on an embodiment of the present application is illustrated, which uses a cloud computing-based preventive maintenance method, a system and an electronic device, which uses a deep neural network model based on cloud computing to predict sealing performance of a sealing ring from environmental information during use of the sealing ring, including but not limited to humidity, temperature, cooling water quality, altitude and the like, to further analyze whether an industrial device having the sealing ring needs preventive maintenance. Therefore, the preventive maintenance prediction is carried out based on the cloud computing and the artificial intelligence technology, so that the effectiveness of the preventive maintenance is effectively improved, the stability of the operation of the industrial equipment is ensured, the frequency and the times of the preventive maintenance can be reduced, and unnecessary preventive maintenance is reduced.
It should be noted that, although the seal ring is exemplified as a seal ring in a hydraulic circuit oil filter on an aircraft in the above embodiments, it should be understood by those skilled in the art that the analysis object of the cloud computing-based preventive maintenance method according to the present application is not limited to a seal ring in a hydraulic circuit oil filter on an aircraft, but is also applicable to a seal ring in a hydraulic circuit oil filter for a vehicle or other environmentally-affected seal rings, and is not limited to the present application.
Exemplary System
Fig. 7 illustrates a block diagram of a cloud-computing-based preventative maintenance system according to an embodiment of the present application. As shown in fig. 7, a cloud computing-based preventative maintenance system 700 according to an embodiment of the present application includes: the environment data acquiring unit 710 is used for acquiring sealing ring performance data of a last routine maintenance test of an airplane at a cloud server based on cloud computing and acquiring environment data of the airplane in a flight process from a data recording device positioned at the end side of the airplane; a performance input vector generating unit 720, configured to pass the seal ring performance data through a hidden markov model to obtain a performance input vector; an environment input vector generating unit 730, configured to pass the environment data through a hidden markov model to obtain an environment input vector; a performance environment feature vector generating unit 740, configured to pass the performance input vector obtained by the performance input vector generating unit 720 and the environment input vector obtained by the environment input vector generating unit 730 through a deep neural network to obtain a current performance feature vector and a current environment feature vector, respectively; a historical feature vector generation unit 750, configured to obtain a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server; a first weighting coefficient generating unit 760 configured to calculate a global Softmax-like function value of the current performance feature vector obtained by the performance input vector generating unit 720 to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponent function values raised by the negative numbers of feature values of the current performance feature vector and the plurality of historical performance feature vectors at respective positions thereof to obtain a plurality of weighted sum values, and then the weighted sum value of the current performance feature vector is divided by the sum of weighted sums of all the current performance feature vectors and the plurality of historical performance feature vectors; a second weighting coefficient generating unit 770 configured to calculate a global Softmax-like function value of the current environment feature vector obtained by the performance environment feature vector generating unit 740 to obtain a second weighting coefficient, where the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponent function values raised by negative numbers of feature values of respective positions of the current environment feature vector and the plurality of historical environment feature vectors, respectively, to obtain a plurality of weighted sum values, and then divide the weighted sum value of the current environment feature vector by a sum of weighted sum values of all the current environment feature vectors and the plurality of historical environment feature vectors; a weighted feature vector generation unit 780 configured to weight the feature value of each position of the current performance feature vector obtained by the performance environment feature vector generation unit 740 with the first weighting coefficient obtained by the first weighting coefficient generation unit 760 to obtain a first weighted vector, and weight the feature value of each position of the current environment feature vector obtained by the performance environment feature vector generation unit 740 with the second weighting coefficient obtained by the second weighting coefficient generation unit 770 to obtain a second weighted feature vector; a classification feature vector generation unit 790 for concatenating the first weighted feature vector obtained by the weighted feature vector generation unit 780 and the second weighted feature vector obtained by the weighted feature vector generation unit 780 to obtain a classification feature vector; and a classification result generating unit 800, configured to pass the classification feature vector obtained by the classification feature vector generating unit 790 through a classifier to obtain a classification result, where the classification result represents a prediction result of a sealing effect of a current sealing ring.
In one example, in the cloud computing-based preventative maintenance system 700, as shown in fig. 8, the performance input vector generation unit 720 includes: a geographic position obtaining subunit 721, configured to obtain geographic position data of the aircraft in a flight process; and a performance input vector composition subunit 722, configured to pass the seal ring performance data through a hidden markov model to obtain the performance input vector based on a correspondence between distances between high-dimensional vectors obtained by conversion of the seal ring performance data and distances between high-dimensional vectors obtained by conversion of the geographical position data.
In one example, in the cloud computing-based preventative maintenance system 700, as shown in fig. 9, the environment input vector generation unit 730 includes: a range distance data acquiring subunit 731, configured to acquire range distance data of the aircraft in a flight process; and an environment input vector configuration subunit 732, configured to pass the environment data through a hidden markov model to obtain the environment input vector based on a correspondence between a distance between high-dimensional vectors obtained by the environment data conversion and a distance between high-dimensional vectors obtained by the voyage distance data conversion.
In one example, in the cloud computing-based preventative maintenance system 700, the first weighting factor generating unit 760 includes: calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient according to the following formula: a1 ═ Σiexp(-xi)/∑j∑iexp (-xi), a1 represents the first weighting coefficient, xi represents the eigenvalue of each position in the current performance eigenvector, and j represents each position of the historical performance eigenvector.
In one example, in the cloud computing-based preventative maintenance system 700, the second weighting factor generating unit 770 includes: calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient according to the following formula: a2 ═ Σiexp(-xi)/∑j∑iexp (-xi), a2 represents the second weighting coefficient, xi represents the feature value of each position in the current environment feature vector, and j represents each position in the historical environment feature vector.
In one example, in the cloud computing-based preventative maintenance system 700, further comprising: storing the current environmental feature vector and the current performance feature vector in the cloud server.
In one example, in the cloud computing-based preventative maintenance system 700, as shown in fig. 10, the classification result generating unit 800 includes: a failure probability generating subunit 801, configured to pass the classification feature vector through a Softmax function to obtain a first probability that a current sealing ring fails and a second probability that the current sealing ring does not fail; and a classification result construction subunit 802 configured to determine the classification result based on the first probability and the second probability.
In one example, in the cloud computing-based preventative maintenance system 700, the deep neural network is a deep fully-connected network.
Here, it can be understood by those skilled in the art that the specific functions and operations of the respective units and modules in the cloud computing-based preventative maintenance system 700 described above have been described in detail in the description of the cloud computing-based preventative maintenance method above with reference to fig. 1 to 6, and thus, a repetitive description thereof will be omitted.
As described above, the cloud computing-based preventative maintenance system 700 according to the embodiment of the present application may be implemented in various terminal devices, such as a server applied to a cloud computing-based preventative maintenance algorithm, and the like. In one example, the cloud computing-based preventative maintenance system 700 according to an embodiment of the present application may be integrated into a terminal device as one software module and/or hardware module. For example, the cloud-based preventative maintenance system 700 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the cloud computing-based preventative maintenance system 700 may also be one of many hardware modules of the terminal device.
Alternatively, in another example, the cloud computing-based preventative maintenance system 700 and the terminal device may also be separate devices, and the cloud computing-based preventative maintenance system 700 may be connected to the terminal device through a wired and/or wireless network and transmit the interaction information in an agreed data format.
Exemplary electronic device
Next, an electronic apparatus according to an embodiment of the present application is described with reference to fig. 11.
As shown in fig. 11, the electronic device 10 includes at least one processor 11 and at least one memory 12.
The processor 11 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.
The memory 12 may include at least one computer program product that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), hard disk, flash memory, etc. At least one computer program instruction may be stored on the computer readable storage medium and executed by the processor 11 to implement the cloud computing-based preventative maintenance methods of the various embodiments of the present application described above and/or other desired functionality. Various contents such as the first weighted feature vector, the class feature vector, etc. may also be stored in the computer readable storage medium.
In one example, the electronic device 10 may further include: an input device 13 and an output device 14, which are interconnected by a bus system and/or other form of connection mechanism (not shown).
The input device 13 may include, for example, a keyboard, a mouse, and the like.
The output device 14 can output various information including the classification result to the outside. The output devices 14 may include, for example, a display, speakers, a printer, and a communication network and its connected remote output devices, among others.
Of course, for the sake of simplicity, only some of the components of the electronic device 10 relevant to the present application are shown in fig. 11, and components such as buses, input/output interfaces, and the like are omitted. In addition, the electronic device 10 may include any other suitable components depending on the particular application.

Claims (10)

1. A preventive maintenance method based on cloud computing is characterized by comprising the following steps:
acquiring sealing ring performance data of a last routine maintenance test of an airplane at a cloud server for cloud computing and acquiring environmental data of the airplane in the flying process from a data recording device positioned at the end side of the airplane;
passing the seal ring performance data through a hidden Markov model to obtain a performance input vector;
passing the environment data through a hidden Markov model to obtain an environment input vector;
respectively passing the performance input vector and the environment input vector through a deep neural network to obtain a current performance characteristic vector and a current environment characteristic vector;
obtaining a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server;
calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient, wherein the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative of the feature values of the current performance feature vector and the plurality of historical performance feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current performance feature vector by the sum of all the current performance feature vectors and the weighted sum values of the plurality of historical performance feature vectors;
calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient, wherein the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponential function values powered by the negative number of the feature value of each position of the current environment feature vector and the plurality of historical environment feature vectors to obtain a plurality of weighted sum values, and dividing the weighted sum value of the current environment feature vector by the sum of all the weighted sum values of the current environment feature vector and the plurality of historical environment feature vectors;
weighting the feature values of the positions of the current performance feature vector by the first weighting coefficient to obtain a first weighted vector, and weighting the feature values of the positions of the current environment feature vector by the second weighting coefficient to obtain a second weighted feature vector;
concatenating the first weighted feature vector and the second weighted feature vector to obtain a classified feature vector; and
and passing the classified feature vectors through a classifier to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
2. The cloud computing-based preventative maintenance method of claim 1, wherein passing the seal ring performance data through a hidden markov model to obtain a performance input vector comprises:
acquiring geographic position data of the airplane in the flight process; and
and passing the sealing ring performance data through a hidden Markov model to obtain the performance input vector based on the corresponding relation between the distance between the high-dimensional vectors obtained by converting the sealing ring performance data and the distance between the high-dimensional vectors obtained by converting the geographical position data.
3. The cloud-computing-based preventative maintenance method of claim 1, wherein passing the environmental data through a hidden markov model to obtain an environmental input vector comprises:
acquiring range distance data of the airplane in the flying process; and
and passing the environment data through a hidden Markov model to obtain the environment input vector based on the corresponding relation between the distance between the high-dimensional vectors obtained by the environment data conversion and the distance between the high-dimensional vectors obtained by the voyage distance data conversion.
4. The cloud computing-based preventative maintenance method of claim 1, wherein computing the global Softmax-like function value for the current performance feature vector to obtain a first weighting coefficient comprises:
calculating a global Softmax-like function value of the current performance feature vector to obtain a first weighting coefficient according to the following formula: a1 ═ Σiexp(-xi)/∑j∑iexp (-xi), a1 represents the first weighting coefficient, xi representsAnd j represents each position of the historical performance feature vector.
5. The cloud computing-based preventative maintenance method of claim 1, wherein computing the global Softmax-like function value for the current environmental feature vector to obtain a second weighting coefficient comprises:
calculating a global Softmax-like function value of the current environment feature vector to obtain a second weighting coefficient according to the following formula: a2 ═ Σiexp(-xi)/∑j∑iexp (-xi), a2 represents the second weighting coefficient, xi represents the feature value of each position in the current environment feature vector, and j represents each position in the historical environment feature vector.
6. The cloud computing-based preventative maintenance method according to claim 1, further comprising: storing the current environmental feature vector and the current performance feature vector in the cloud server.
7. The cloud-computing-based preventative maintenance method of claim 1, wherein passing the classified feature vectors through a classifier to obtain classification results comprises:
the classification characteristic vector is processed by a Softmax function to obtain a first probability that the current sealing ring fails and a second probability that the current sealing ring does not fail; and
determining the classification result based on the first probability and the second probability.
8. The cloud computing-based preventative maintenance method of claim 1, wherein the deep neural network is a deep fully-connected network.
9. A cloud computing-based preventative maintenance system, comprising:
the device comprises an environmental data acquisition unit, a data recording unit and a data processing unit, wherein the environmental data acquisition unit is used for acquiring sealing ring performance data of the last routine maintenance test of the airplane at a cloud server based on cloud computing and acquiring environmental data of the airplane in the flying process from a data recording device positioned at the end side of the airplane;
the performance input vector generating unit is used for enabling the sealing ring performance data to pass through a hidden Markov model so as to obtain a performance input vector;
the environment input vector generating unit is used for enabling the environment data obtained by the environment data obtaining unit to pass through a hidden Markov model so as to obtain an environment input vector;
a performance environment feature vector generation unit, configured to pass the performance input vector obtained by the performance input vector generation unit and the environment input vector obtained by the environment input vector generation unit through a deep neural network, respectively, to obtain a current performance feature vector and a current environment feature vector;
a historical feature vector generation unit, configured to obtain a plurality of historical performance feature vectors and a plurality of historical environment feature vectors from the cloud server;
a first weighting coefficient generation unit configured to calculate a global Softmax-like function value of the current performance feature vector obtained by the performance input vector generation unit to obtain a first weighting coefficient, where the global Softmax-like function value of the current performance feature vector is equal to a weighted sum of natural constant exponent function values raised by negatives of feature values of respective positions of the current performance feature vector and the plurality of historical performance feature vectors, respectively, to obtain a plurality of weighted sum values, and the weighted sum value of the current performance feature vector is divided by a sum of weighted sum values of all of the current performance feature vectors and the plurality of historical performance feature vectors;
a second weighting coefficient generation unit configured to calculate a global Softmax-like function value of the current environment feature vector obtained by the performance environment feature vector generation unit to obtain a second weighting coefficient, where the global Softmax-like function value of the current environment feature vector is equal to a weighted sum of natural constant exponent function values raised by negatives of feature values of respective positions of the current environment feature vector and the plurality of historical environment feature vectors, respectively, to obtain a plurality of weighted sum values, and then divide the weighted sum value of the current environment feature vector by a sum of weighted sum values of all the current environment feature vectors and the plurality of historical environment feature vectors;
a weighted feature vector generation unit configured to weight the feature value of each position of the current performance feature vector obtained by the performance environment feature vector generation unit with the first weighting coefficient obtained by the first weighting coefficient generation unit to obtain a first weighted vector, and weight the feature value of each position of the current performance feature vector obtained by the performance environment feature vector generation unit with the second weighting coefficient obtained by the second weighting coefficient generation unit to obtain a second weighted feature vector;
a classification feature vector generation unit configured to cascade the first weighted feature vector obtained by the weighted feature vector generation unit and the second weighted feature vector obtained by the weighted feature vector generation unit to obtain a classification feature vector; and
and the classification result generating unit is used for enabling the classification characteristic vector obtained by the classification characteristic vector generating unit to pass through a classifier so as to obtain a classification result, wherein the classification result represents a prediction result of the sealing effect of the current sealing ring.
10. An electronic device, comprising:
a processor; and
a memory having stored therein computer program instructions that, when executed by the processor, cause the processor to perform the cloud computing-based preventative maintenance method of any one of claims 1-8.
CN202110530260.XA 2021-05-14 2021-05-14 Preventive maintenance method based on cloud computing Active CN113158346B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110530260.XA CN113158346B (en) 2021-05-14 2021-05-14 Preventive maintenance method based on cloud computing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110530260.XA CN113158346B (en) 2021-05-14 2021-05-14 Preventive maintenance method based on cloud computing

Publications (2)

Publication Number Publication Date
CN113158346A true CN113158346A (en) 2021-07-23
CN113158346B CN113158346B (en) 2022-08-09

Family

ID=76876095

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110530260.XA Active CN113158346B (en) 2021-05-14 2021-05-14 Preventive maintenance method based on cloud computing

Country Status (1)

Country Link
CN (1) CN113158346B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752276A (en) * 2012-02-02 2012-10-24 青岛印象派信息技术有限公司 Verification code service method and system based on cloud computing
US20190095265A1 (en) * 2017-09-26 2019-03-28 At&T Intellectual Property I, L.P. Intelligent Preventative Maintenance of Critical Applications in Cloud Environments
CN109948169A (en) * 2017-12-20 2019-06-28 中国中车股份有限公司 A kind of railway freight-car prognostic and health management system
US10445673B1 (en) * 2018-03-15 2019-10-15 American Public Power Assocation Inc. Predictive models for electric power grid maintenance
EP3591530A1 (en) * 2018-07-02 2020-01-08 Accenture Global Solutions Limited Intelligent backup and recovery of cloud computing environment
CN112100773A (en) * 2020-09-14 2020-12-18 电子科技大学 Turbine machinery preventive maintenance method based on component correlation
CN112329151A (en) * 2020-11-19 2021-02-05 中国航空工业集团公司沈阳飞机设计研究所 Preventive maintenance method for components
CN112766810A (en) * 2021-02-04 2021-05-07 广州蕊生网络科技有限公司 Neural network training method for intelligent comprehensive overall quality evaluation

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752276A (en) * 2012-02-02 2012-10-24 青岛印象派信息技术有限公司 Verification code service method and system based on cloud computing
US20190095265A1 (en) * 2017-09-26 2019-03-28 At&T Intellectual Property I, L.P. Intelligent Preventative Maintenance of Critical Applications in Cloud Environments
CN109948169A (en) * 2017-12-20 2019-06-28 中国中车股份有限公司 A kind of railway freight-car prognostic and health management system
US10445673B1 (en) * 2018-03-15 2019-10-15 American Public Power Assocation Inc. Predictive models for electric power grid maintenance
EP3591530A1 (en) * 2018-07-02 2020-01-08 Accenture Global Solutions Limited Intelligent backup and recovery of cloud computing environment
CN112100773A (en) * 2020-09-14 2020-12-18 电子科技大学 Turbine machinery preventive maintenance method based on component correlation
CN112329151A (en) * 2020-11-19 2021-02-05 中国航空工业集团公司沈阳飞机设计研究所 Preventive maintenance method for components
CN112766810A (en) * 2021-02-04 2021-05-07 广州蕊生网络科技有限公司 Neural network training method for intelligent comprehensive overall quality evaluation

Also Published As

Publication number Publication date
CN113158346B (en) 2022-08-09

Similar Documents

Publication Publication Date Title
Lim et al. A time window neural network based framework for remaining useful life estimation
Tong et al. A novel deep learning method for aircraft landing speed prediction based on cloud-based sensor data
US20050143962A1 (en) Computational design methods
Jee Keen Raymond et al. Classification of partial discharge measured under different levels of noise contamination
Mukherjee et al. Development of an ensemble decision tree-based power system dynamic security state predictor
Sahani et al. Deep convolutional stack autoencoder of process adaptive VMD data with robust multikernel RVFLN for power quality events recognition
CN116577677A (en) Discharging test system and method for retired power battery
Rodríguez‐Borbón et al. Reliability estimation for accelerated life tests based on a Cox proportional hazard model with error effect
CN116451139A (en) Live broadcast data rapid analysis method based on artificial intelligence
CN111276247A (en) Flight parameter data health assessment method and device based on big data processing and readable storage medium
CN113158346B (en) Preventive maintenance method based on cloud computing
Xia et al. Research on forecasting aeroengine vibration signals based on the MAE model
Araguz et al. Optimized model-based design space exploration of distributed multi-orbit multi-platform Earth observation spacecraft architectures
Mahmoud et al. Different long short-term memory approaches to enhance prediction-based satellite telemetry compression
Wade et al. Machine learning algorithms for hums improvement on rotorcraft components
CN110988631A (en) Partial discharge positioning detection method and device, readable storage medium and electrical equipment
Romankevich et al. Fault-tolerant multiprocessor systems reliability estimation using statistical experiments with GL-models
Leal-Romo et al. Selecting surrogate-based modeling techniques for power integrity analysis
CN115456048A (en) Reliability evaluation system and method for offshore wind turbine
Zheng et al. Discriminant generative adversarial networks with its application to equipment health classification
Li et al. A transformer fault diagnosis method based on sub-clustering reduction and multiclass multi-kernel support vector machine
Tang et al. Predicting aircraft taxiing estimated time of arrival by cluster analysis
Leao et al. Asymmetric unscented transform for failure prognosis
Mesquita et al. Detection and fault prediction in electrolytic capacitors using artificial neural networks
Garreta Piñol Artificial intelligence for aircraft predictive maintenance

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
TA01 Transfer of patent application right

Effective date of registration: 20220715

Address after: 271000 No.1, area a, Taishan science and technology venture City, 1110 Nantianmen street, high tech Zone, Tai'an City, Shandong Province

Applicant after: SHANDONG ZHONGZHI ELECTRONICS CO.,LTD.

Address before: No.583 Jinqiao street, Qiantang New District, Hangzhou, Zhejiang 310000

Applicant before: Hangzhou Jiying Technology Co.,Ltd.

TA01 Transfer of patent application right
GR01 Patent grant
GR01 Patent grant