CN110502724B

CN110502724B - Equipment state prediction method based on self-organizing neural network and terminal equipment

Info

Publication number: CN110502724B
Application number: CN201910732373.0A
Authority: CN
Inventors: 边伟; 贺卫华; 刘国瑞; 刘雷
Original assignee: State Grid Shanxi Electric Power Co Ltd
Current assignee: State Grid Shanxi Electric Power Co Ltd
Priority date: 2019-08-09
Filing date: 2019-08-09
Publication date: 2023-04-07
Anticipated expiration: 2039-08-09
Also published as: CN110502724A

Abstract

The application is applicable to the technical field of power automation, and provides a device state prediction method and terminal device based on a self-organizing neural network, wherein the method comprises the following steps: acquiring original telemetering data, and extracting corresponding typical defects according to the original telemetering data; calculating the actual period of the typical defects and the weight of key telemeasurement, establishing a regression model corresponding to the typical defects, and predicting the equipment state of the typical defects. The self-organizing neural network-based equipment state prediction method provided by the embodiment of the application can identify the key remote measurement which is difficult to perceive and does not exceed the guiding rule or the regulation limit value when the equipment is in the initial stage of the abnormal state, solves the problem that the fault and latent fault of the power transmission and transformation equipment in the initial stage of insulation degradation cannot be identified in time in the prior art, and is beneficial to improving the operation reliability of the power transmission and transformation equipment.

Description

Equipment state prediction method based on self-organizing neural network and terminal equipment

Technical Field

The application belongs to the technical field of electric power automation, and particularly relates to a device state prediction method based on a self-organizing neural network and a terminal device.

Background

The measured data changes little and regularly in the normal operation process of the power grid, belongs to a stable sequence, and can be described by adopting an auto-regressive (AR) model. The other part of the state quantity changes day-to-day periodically, but the change amplitude is not large, such as oil temperature, environment temperature and the like, and the state quantity also conforms to the AR model after the periodicity is removed. Therefore, a time series autoregressive model is adopted to represent the change of the single state quantity data with time.

Because the insulation degradation process or the latent fault of the power transmission and transformation equipment is slow to develop, when the equipment is in an abnormal state, the associated measurement data often does not exceed the limit value in the guiding rule or the regulation, so that the associated measurement data is difficult to perceive. Therefore, for the online monitoring data which does not exceed the state quantity limit value, the measurement data jump caused by the fault cannot be identified by simply adopting the AR model.

Disclosure of Invention

In view of this, the embodiment of the present application provides a device state prediction method and a terminal device based on a self-organizing neural network, so as to solve the problem that the current power transmission and transformation device monitoring technology cannot identify a fault at an initial stage of insulation degradation and a latent fault.

According to a first aspect, an embodiment of the present application provides a device state prediction method based on an ad hoc neural network, including: acquiring original telemetering data, and extracting corresponding typical defects according to the original telemetering data; calculating the actual period of the typical defects and each remote measurement of abnormal change in the actual period of the typical defects according to the original remote measuring data and the typical defects; recording each telemetered quantity which is abnormally changed in the actual period of the typical defect as a key telemetered quantity; calculating the weight of each key telemetering measurement corresponding to the typical defect according to the original telemetering data; establishing a regression model corresponding to the typical defects according to the actual period of the typical defects and the weight of each key telemeasurement corresponding to the typical defects; and predicting the equipment state of the typical defect according to the regression model corresponding to the typical defect.

With reference to the first aspect, in some embodiments of the present application, the extracting the corresponding typical defect from the raw telemetry data includes: performing text analysis on the original telemetering data to obtain a corresponding keyword set; clustering each keyword in the keyword set, and determining the typical defects according to the accumulated result.

With reference to the first aspect, in some embodiments of the present application, the calculating, according to the raw telemetry data and the typical defect, an actual period of the typical defect and each telemetry measure of abnormal changes in the actual period of the typical defect includes: acquiring a transition probability matrix corresponding to the telemetry data of the original telemetry data in the normal state of the historical equipment according to a preset self-organizing neural network model; acquiring the recording start time of the typical defect according to the original telemetering data, and counting abnormal points before the recording start time and the abnormal time corresponding to each abnormal point; recording the earliest time in the variation time corresponding to each abnormal point as the defect starting time; acquiring defect elimination time of the typical defect according to the original telemetry data; recording the defect eliminating time as defect ending time; calculating an actual period of occurrence of the typical defect according to the defect start time and the defect end time; drawing a defect characteristic curve of each telemetering amount changing along with time in the actual period of the typical defect; and determining each remote measurement which is abnormally changed in the actual period of the typical defect according to the defect characteristic curve of each remote measurement along with the change of time.

With reference to the first aspect, in some embodiments of the present application, the calculating, from the raw telemetry data, a weight of each key telemetry measure corresponding to the typical defect includes: calculating a transition probability matrix of any key telemetering measurement in a normal state; calculating a transition probability matrix of any key telemetering measurement in an abnormal state according to the actual period of the typical defect; and calculating the weight of any key telemetering measurement according to the transition probability matrix of any key telemetering measurement in the normal state and the transition probability matrix of any key telemetering measurement in the abnormal state.

With reference to the first aspect, in some embodiments of the present application, the calculating a weight of any key telemetry measure according to a transition probability matrix of the any key telemetry measure in a normal state and a transition probability matrix of the any key telemetry measure in an abnormal state includes: calculating a difference absolute value and a corresponding confidence interval according to the transition probability matrix of any key telemetering measurement in a normal state and the transition probability matrix of any key telemetering measurement in an abnormal state; multiplying each difference in the confidence interval by the corresponding posterior probability to calculate a corresponding expected value; the expected value is a weight of the any key telemetry measurement.

With reference to the first aspect, in some embodiments of the present application, establishing a regression model corresponding to the typical defect according to the actual period of the typical defect and the weight of each key telemetering measurement corresponding to the typical defect includes: and according to a regression analysis algorithm, obtaining a regression model corresponding to the typical defects.

With reference to the first aspect, in some embodiments of the present application, the predicting the device state of the typical defect according to the regression model corresponding to the typical defect includes: substituting the measured data of each key telemeasurement corresponding to the typical defect into the regression model, and calculating to obtain a corresponding predicted value; and when the predicted value is greater than or equal to a preset threshold value, judging that the equipment state is abnormal, and determining that the corresponding abnormal type is the typical defect.

According to a second aspect, an embodiment of the present application provides a terminal device, including: the input end member is used for acquiring original telemetering data and extracting corresponding typical defects according to the original telemetering data; the first calculation unit is used for calculating the actual period of the typical defect and each telemeasurement of abnormal change in the actual period of the typical defect according to the original telemetering data and the typical defect; recording each telemetered quantity which is abnormally changed in the actual period of the typical defect as a key telemetered quantity; the second calculation unit is used for calculating the weight of each key telemetering measurement corresponding to the typical defect according to the original telemetering data; the model establishing unit is used for establishing a regression model corresponding to the typical defect according to the actual period of the typical defect and the weight of each key telemeasurement corresponding to the typical defect; and the prediction unit is used for predicting the equipment state of the typical defect according to the regression model corresponding to the typical defect.

According to a third aspect, an embodiment of the present application provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the steps of the method according to the first aspect or any implementation manner of the first aspect.

According to a fourth aspect, embodiments of the present application provide a computer-readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the steps of the method according to the first aspect or any embodiment of the first aspect.

The self-organizing neural network-based equipment state prediction method provided by the embodiment of the application can be used for identifying key telemeasurement which is difficult to detect and does not exceed a guide rule or a regulation limit value when the equipment is in an abnormal state initial stage aiming at the condition that the insulation degradation process or the latent fault of the power transmission and transformation equipment develops slowly, so that the jumping identification of fault data which does not exceed the state quantity limit value is realized, the problem that the fault and the latent fault of the power transmission and transformation equipment in the insulation degradation initial stage cannot be identified in time in the prior art is solved, and the reliability of the operation of the power transmission and transformation equipment is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

Fig. 1 is a schematic flowchart of a method for predicting a device state based on an ad hoc neural network according to an embodiment of the present application;

fig. 2 is a schematic structural diagram of a terminal device according to an embodiment of the present application;

fig. 3 is a schematic structural diagram of another terminal device according to an embodiment of the present application.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

Reliable power equipment is the basis for ensuring the safe operation of a power grid and improving the reliability of power supply. In actual operation, some power equipment can continue to be used, but the operation state is abnormal or hidden troubles exist, and personal safety, reliable and economic operation of a power grid, equipment service life, power quality and the like are influenced. Such anomalies or potential hazards are referred to as defects. In the operation and maintenance process of the power system, the defects of the power equipment can jeopardize the safe and stable operation of the whole system. The potential safety hazard of the equipment can be timely discovered by predicting the possibility of future defects in the operation process of the power system, so that the safe and controllable operation of the equipment is realized. The equipment defect algorithm flow based on big data mining is as follows:

firstly, finding out the telemetering data characteristics of the equipment when the defect occurs according to the defect of the equipment and the telemetering amount historical data when the defect occurs, then comparing the telemetering data characteristics of a future time period predicted by regression analysis with the defect data characteristics of the history occurrence, and finding out the similarity of the telemetering data characteristics and the defect data characteristics, thereby quantitatively evaluating the possibility of the future occurrence of similar defects. The algorithm is mainly divided into four key links: typical defect analysis, key telemetering amount searching and weight analysis, telemetering amount variation trend analysis and equipment defect prediction.

In order to explain the technical means described in the present application, the following description will be given by way of specific examples.

An embodiment of the present application provides a device state prediction method based on a self-organizing neural network, and as shown in fig. 1, the device state prediction method based on the self-organizing neural network may include the following steps:

step S101: raw telemetry data is acquired, and corresponding typical defects are extracted according to the raw telemetry data.

In a specific real-time mode, the original telemetered data is subjected to text analysis to obtain a corresponding keyword set, and then each keyword in the keyword set is clustered, and typical defects are determined according to the accumulated results.

When different types of defects occur, the data characteristics of the accompanying telemetry data change are often different, so that when the defects are predicted, the defects need to be analyzed respectively according to the types of the defects. And the defect description in the current defect history record is often described by an unstructured natural language. Therefore, firstly, a text parsing technology is adopted to convert unstructured defect descriptions into structured languages, and then the structured defect descriptions are spoken for clustering, so that defects with more defect instances in a cluster are found out, and typical defects are obtained by combining service analysis.

Text clustering refers to the process of grouping text clusters into a set consisting of several text clusters. Text clustering generally involves two steps:

and (6) text parsing. The text is unstructured data, before analysis, the representation of the text and the selection of characteristic items thereof are needed, characteristic words extracted from the text are converted into a structured data form from an unstructured original text, firstly, word segmentation processing, word sequence induction, non-keyword and language auxiliary words are filtered, so that a keyword set of the text is obtained, then, a model is used for conversion, the text is represented into a form which can be processed by a computer by a simple and accurate method, namely, text characteristic expression, a probability model method is adopted in the text, the method comprehensively considers factors such as word frequency, document length and the like, documents and user interests are fused according to a certain probability relation, semantic similarity of the two texts is measured through probability in a probability measurement space, and the method is successfully applied in the field of information retrieval.

For example, actual defect information in the power grid, that is, "main transformer No. 3 5041 switches C-phase SF6 low pressure. The nominal pressure was 0.7MPA and the field pressure was 0.63MPA. After the two steps, the phases can be divided into a main transformer 5041 No. 3 switch C phase representing equipment name, SF6 low pressure representing defect type and rated pressure 0.7MPA representing defect description, and the site pressure is 0.63MPA;

and clustering texts. And clustering the characteristic texts by using a text clustering method based on a vector space model, wherein vectors with large quantity are typical defects. The vector space model was proposed by Gerard Salton et al in the late 60 s, was successfully applied to the well-known SMART text retrieval system, and was later widely applied in the field of text clustering. The main idea of this model is to map each text to a point in a vector space consisting of a set of normalized orthogonal feature vectors. Each dimension of the vector corresponds to a characteristic item in the text, the characteristic item refers to various units forming the text, such as words, phrases and the like, and the weight of the characteristic item reflects the main degree of the characteristic item to the content of the text. For a text, the text is a plurality of different entries, the weights of the characteristic items in the text are recorded, and the text is represented as

If the historical defect records of the 500kV main transformer are analyzed, the defect records with characteristic items of 'oil temperature', 'abnormity', 'high', 'actual value' and 'alarm value' are 312, and account for 21 percent of the total number of samples. Therefore, the oil temperature abnormal defect is a typical defect for a 500kV main transformer.

Step S102: and calculating the actual period of the typical defect and each remote measurement of abnormal change in the actual period of the typical defect according to the original remote measuring data and the typical defect. Each telemetry that abnormally changes during the actual period of a typical defect is recorded as a key telemetry.

In a specific real-time manner, the process of step S102 can be implemented by the following several sub-steps:

step S1021: and acquiring a transition probability matrix corresponding to the telemetry data of the historical equipment in the normal state in the original telemetry data according to a preset self-organizing neural network model.

Step S1022: and acquiring the recording start time of the typical defect according to the original telemetering data, and counting abnormal points before the recording start time and the abnormal time corresponding to each abnormal point. And recording the earliest time in the variation time corresponding to each abnormal point as the defect starting time.

Step S1023: and acquiring the defect elimination time of the typical defect according to the original telemetry data. And recording the defect eliminating time as the defect ending time.

Step S1024: calculating an actual period in which the typical defect occurs according to the defect start time and the defect end time.

Step S1025: and drawing a defect characteristic curve of each telemetric quantity changing along with time in the actual period of the typical defect.

Step S1026: and determining each remote measurement which is abnormally changed in the actual period of the typical defect according to the defect characteristic curve of each remote measurement along with the change of time.

To predict the similarity between the future telemetry variation trend and the historical defect occurrence, it is first necessary to find the actual period of defect occurrence and the telemetry amount that abnormally varies in the period. Meanwhile, as part of latent defects of the power equipment are slowly developed, the patrol personnel are often difficult to perceive when the equipment is just in an abnormal state, and the actual occurrence time of the defects is often earlier than the time recorded in the defect history record. Therefore, the period of occurrence of the defect needs to be analyzed to obtain the time for finding out the occurrence of the whole defect, so that the data characteristics of the relevant remote measurement in the period of occurrence of the defect can be obtained.

Firstly, obtaining a data change rule of the telemetering measurement in a normal state, namely a transition probability matrix of the telemetering measurement in the normal state based on the telemetering data of the historical equipment in the normal state according to a regression model and a self-organizing neural network algorithm. Comparing the starting time of defect record to be detected and all measured values of the telemetering data one day before the starting time with the data change rule of the telemetering data in the normal state, wherein the data change rule is not met, namely the point with the transition probability of 0 is an abnormal point, and the time corresponding to the point is the time when the telemetering data is abnormal. And obtaining the time of the variability of each telemetering item associated with the defect when the defect occurs according to the judgment process, taking the earliest time of the variability as the starting time of the defect, taking the defect elimination time in the defect record as the ending time of the defect, thereby obtaining the time period of each defect, and obtaining the defect characteristic curve of each telemetering quantity changing along with the time when the defect occurs according to the time period.

Step S103: weights for each of the key telemetry measurements corresponding to the representative defect are calculated from the raw telemetry data.

In a specific real-time manner, the process of step S103 can be implemented by the following sub-steps:

step S1031: calculating a transition probability matrix of any key telemetering measurement in a normal state;

step S1032: calculating a transition probability matrix of any key telemetering measurement in an abnormal state according to the actual period of the typical defect;

step S1033: and calculating the weight of any key telemetering measurement according to the transition probability matrix of any key telemetering measurement in the normal state and the transition probability matrix of any key telemetering measurement in the abnormal state.

Specifically, the absolute value of the difference and the corresponding confidence interval may be calculated according to the transition probability matrix of any key telemetry measurement in the normal state and the transition probability matrix of any key telemetry measurement in the abnormal state; and multiplying each difference in the confidence interval by the corresponding posterior probability to calculate a corresponding expected value, wherein the expected value is the weight of any key telemeasurement.

The purpose of the algorithm is to calculate the magnitude of the abnormal change of the telemetering amount of the defect when the defect occurs as the weight in the similarity analysis. The difference between the abnormal change amplitude and the normal data regularity cannot be evaluated only by considering the abnormal change amplitude in the abnormal data state. Therefore, the amplitude of the abnormal change of the telemetering amount is represented by comparing the data change rule in the normal state with the data change rule in the abnormal state and calculating the expected value of the difference value.

The algorithm idea is as follows: firstly, obtaining the data change rule of the telemetering measurement in the normal state, namely a transition probability matrix in the normal state. And similarly, a data change rule of the telemetering amount in the abnormal state, namely a transition probability matrix in the abnormal state is obtained based on the telemetering data in the defect occurrence period.

Based on Bayes theory, by calculating the most probable value and probability (maximum posterior distribution) of each time point when the defect occurs in the remote measurement, formula is used

/>

And calculating a probability transition matrix of the remote measurement in the defect occurrence time period in the normal state. And comparing the probability transition matrix with the obtained abnormal data probability transition matrix when the defect occurs. And calculating the difference absolute value of the state value corresponding to each telemetering variable and the confidence interval of the difference according to the two transition probability matrixes. Each difference in the confidence interval is multiplied by its corresponding probability of occurrence (i.e., a posterior probability) to calculate its expected value, which represents the magnitude of the abnormal change in telemetry.

Step S104: and establishing a regression model corresponding to the typical defects according to the actual period of the typical defects and the weight of each key telemeasurement corresponding to the typical defects.

Specifically, a regression model corresponding to the typical defect may be established according to a regression analysis algorithm.

After determining which telemetering quantities change abnormally in the defect occurrence period, the change trend of the telemetering quantities can be analyzed according to the telemetering quantities to obtain the function of the change of the telemetering quantities along with time, so that the change trend of the telemetering quantities in the future time period can be predicted. Regression analysis is an analysis method which utilizes the data statistical principle to mathematically process a large amount of statistical data, determines the correlation between dependent variables and some independent variables, and establishes a regression equation (function expression) with better correlation for predicting the change of the dependent variables in the future. And analyzing the variation trend of related influence factors of the typical defect of the abnormal main transformer oil temperature by adopting a regression analysis method, and further obtaining a prediction function of each factor.

Step S105: and predicting the equipment state of the typical defect according to a regression model corresponding to the typical defect.

The measured data of each key telemetering measurement corresponding to the typical defect can be substituted into the regression model, and a corresponding predicted value is obtained through calculation. And when the predicted value is greater than or equal to a preset threshold value, judging that the equipment state is abnormal, and determining that the corresponding abnormal type is the typical defect.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by functions and internal logic of the process, and should not constitute any limitation to the implementation process of the embodiments of the present application.

An embodiment of the present application further provides a terminal device, as shown in fig. 2, where the terminal device may include: an input end element 201, a first calculation unit 202, a second calculation unit 203, a model building unit 204 and a prediction unit 205.

The input end member 201 is used for acquiring original telemetering data and extracting corresponding typical defects according to the original telemetering data; the corresponding working process can be referred to step S101 in the above method embodiment.

The first calculating unit 202 is configured to calculate, according to the raw telemetry data and the typical defect, an actual period of the typical defect and each telemetry measure of abnormal changes occurring in the actual period of the typical defect; recording each telemetered quantity which is abnormally changed in the actual period of the typical defect as a key telemetered quantity; the corresponding working process can be referred to step S102 in the above method embodiment.

The second calculating unit 203 is configured to calculate, according to the raw telemetry data, weights of the key telemetry measurements corresponding to the typical defects; the corresponding working process can be referred to step S103 in the above method embodiment.

The model establishing unit 204 is configured to establish a regression model corresponding to the typical defect according to the actual period of the typical defect and the weight of each key telemetering measurement corresponding to the typical defect; the corresponding working process can be referred to as step S104 in the above method embodiment.

The prediction unit 205 is configured to perform equipment state prediction on the typical defect according to a regression model corresponding to the typical defect; the corresponding working process can be referred to step S105 in the above method embodiment.

Fig. 3 is a schematic diagram of a terminal device according to an embodiment of the present application. As shown in fig. 3, the terminal device 600 of this embodiment includes: a processor 601, a memory 602, and a computer program 603, such as a cluster-based device operational state assessment program, stored in the memory 602 and executable on the processor 601. The processor 601, when executing the computer program 603, implements the steps in each of the embodiments of the self-organizing neural network-based device state prediction method described above, such as the steps S101 to S105 shown in fig. 1. Alternatively, the processor 601, when executing the computer program 603, implements the functions of each module/unit in each device embodiment described above, such as the functions of the input unit 201, the first classification unit 202, and the second classification unit 203 shown in fig. 2.

The computer program 603 may be partitioned into one or more modules/units that are stored in the memory 602 and executed by the processor 601 to complete the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing certain functions, which are used to describe the execution of the computer program 603 in the terminal device 600. For example, the computer program 603 may be partitioned into a synchronization module, a summarization module, an acquisition module, a return module (a module in a virtual device).

The terminal device 600 may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The terminal device may include, but is not limited to, a processor 601, a memory 602. Those skilled in the art will appreciate that fig. 3 is merely an example of a terminal device 600 and does not constitute a limitation of terminal device 600 and may include more or fewer components than shown, or some components may be combined, or different components, e.g., the terminal device may also include input-output devices, network access devices, buses, etc.

The Processor 601 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 602 may be an internal storage unit of the terminal device 600, such as a hard disk or a memory of the terminal device 600. The memory 602 may also be an external storage device of the terminal device 600, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the terminal device 600. Further, the memory 602 may also include both an internal storage unit and an external storage device of the terminal apparatus 600. The memory 602 is used for storing the computer programs and other programs and data required by the terminal device. The memory 602 may also be used to temporarily store data that has been output or is to be output.

It should be clear to those skilled in the art that, for convenience and simplicity of description, the foregoing division of the functional units and modules is only used for illustration, and in practical applications, the above function distribution may be performed by different functional units and modules as needed, that is, the internal structure of the apparatus may be divided into different functional units or modules to perform all or part of the above described functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other ways. For example, the above-described embodiments of the apparatus/terminal device are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or may also be implemented in the form of a software functional unit.

The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow in the method of the embodiments described above can be realized by a computer program, which can be stored in a computer-readable storage medium and can realize the steps of the embodiments of the methods described above when the computer program is executed by a processor. . Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U.S. disk, removable hard disk, magnetic diskette, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signal, telecommunications signal, and software distribution medium, etc. It should be noted that the computer-readable medium may contain suitable additions or subtractions depending on the requirements of legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer-readable media may not include electrical carrier signals or telecommunication signals in accordance with legislation and patent practice.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the embodiments of the present application, and they should be construed as being included in the present application.

Claims

1. A device state prediction method based on a self-organizing neural network is characterized by comprising the following steps:

acquiring original telemetering data, and extracting corresponding typical defects according to the original telemetering data;

calculating the actual period of the typical defects and each remote measurement of abnormal change in the actual period of the typical defects according to the original remote measuring data and the typical defects; the method comprises the following steps: acquiring a transition probability matrix corresponding to the telemetry data of the historical equipment in a normal state in the original telemetry data according to a preset self-organizing neural network model; acquiring the recording start time of the typical defect according to the original telemetering data, and counting abnormal points before the recording start time and abnormal time corresponding to each abnormal point; recording the earliest time in the variation time corresponding to each abnormal point as the defect starting time; acquiring defect elimination time of the typical defect according to the original telemetry data; recording the defect eliminating time as defect ending time; calculating an actual period of occurrence of the typical defect according to the defect start time and the defect end time; drawing a defect characteristic curve of each telemetering amount changing along with time in the actual period of the typical defect; determining each telemetering measurement which is abnormally changed in the actual period of the typical defect according to the defect characteristic curve of each telemetering measurement along with the change of time;

recording each telemetering quantity which is abnormally changed in the actual period of the typical defect as a key telemetering quantity;

calculating the weight of each key telemetering measurement corresponding to the typical defect according to the original telemetering data;

establishing a regression model corresponding to the typical defects according to the actual period of the typical defects and the weight of each key telemeasurement corresponding to the typical defects;

and predicting the equipment state of the typical defect according to the regression model corresponding to the typical defect.

2. The method of self-organizing neural network-based device state prediction of claim 1, wherein the extracting corresponding typical defects from the raw telemetry data comprises:

performing text analysis on the original telemetering data to obtain a corresponding keyword set;

clustering each keyword in the keyword set, and determining the typical defects according to the accumulated result.

3. The self-organizing neural network-based device state prediction method of claim 2, wherein the calculating weights of the key telemetry measures corresponding to the typical defect according to the raw telemetry data comprises:

calculating a transition probability matrix of any key telemetering measurement in a normal state;

calculating a transition probability matrix of any key telemetering measurement in an abnormal state according to the actual period of the typical defect;

and calculating the weight of any key telemetering measurement according to the transition probability matrix of any key telemetering measurement in the normal state and the transition probability matrix of any key telemetering measurement in the abnormal state.

4. The method according to claim 3, wherein the calculating the weight of any key telemetry measure according to the transition probability matrix of any key telemetry measure in the normal state and the transition probability matrix of any key telemetry measure in the abnormal state comprises:

calculating a difference absolute value and a corresponding confidence interval according to the transition probability matrix of any key telemetering measurement in a normal state and the transition probability matrix of any key telemetering measurement in an abnormal state;

multiplying each difference in the confidence interval by the corresponding posterior probability to calculate a corresponding expected value; the expected value is a weight of the any key telemetry measurement.

5. The method according to claim 4, wherein the establishing a regression model corresponding to the typical defect according to the actual period of the typical defect and the weight of each key telemetering measurement corresponding to the typical defect comprises:

and establishing a regression model corresponding to the typical defects according to a regression analysis algorithm.

6. The self-organizing neural network-based device state prediction method of claim 5, wherein the predicting the device state of the typical defect according to the regression model corresponding to the typical defect comprises:

substituting the measured data of each key telemetering measurement corresponding to the typical defect into the regression model, and calculating to obtain a corresponding predicted value;

and when the predicted value is greater than or equal to a preset threshold value, judging that the equipment state is abnormal, wherein the corresponding abnormal type is the typical defect.

7. A terminal device, comprising:

the input end member is used for acquiring original telemetering data and extracting corresponding typical defects according to the original telemetering data;

the first calculation unit is used for calculating the actual period of the typical defect and each remote measurement of abnormal change in the actual period of the typical defect according to the original telemetering data and the typical defect; the method comprises the following steps: acquiring a transition probability matrix corresponding to the telemetry data of the original telemetry data in the normal state of the historical equipment according to a preset self-organizing neural network model; acquiring the recording start time of the typical defect according to the original telemetering data, and counting abnormal points before the recording start time and abnormal time corresponding to each abnormal point; recording the earliest time in the variation time corresponding to each abnormal point as the defect starting time; acquiring defect elimination time of the typical defect according to the original telemetry data; recording the defect eliminating time as defect ending time; calculating an actual period of occurrence of the typical defect according to the defect start time and the defect end time; drawing a defect characteristic curve of each telemetering amount changing along with time in the actual period of the typical defect; determining each telemetering measurement which is abnormally changed in the actual period of the typical defect according to the defect characteristic curve of each telemetering measurement along with the change of time;

the first computing unit is also used for recording each telemetering quantity which is abnormally changed in the actual period of the typical defect as a key telemetering quantity;

the second calculation unit is used for calculating the weight of each key telemetering measurement corresponding to the typical defect according to the original telemetering data;

the model establishing unit is used for establishing a regression model corresponding to the typical defect according to the actual period of the typical defect and the weight of each key telemeasurement corresponding to the typical defect;

and the prediction unit is used for predicting the equipment state of the typical defect according to the regression model corresponding to the typical defect.

8. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 6.