CN111723519A - Transformer fault diagnosis device and method based on logistic regression and naive Bayes - Google Patents
Transformer fault diagnosis device and method based on logistic regression and naive Bayes Download PDFInfo
- Publication number
- CN111723519A CN111723519A CN202010472929.XA CN202010472929A CN111723519A CN 111723519 A CN111723519 A CN 111723519A CN 202010472929 A CN202010472929 A CN 202010472929A CN 111723519 A CN111723519 A CN 111723519A
- Authority
- CN
- China
- Prior art keywords
- data
- transformer fault
- machine learning
- prediction
- logistic regression
- 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
Links
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention discloses a transformer fault diagnosis device and method based on logistic regression and naive Bayes. The training result model packaged by the algorithm is packaged in the module in a hardware form, and the multi-core CPU is selected as the hardware support, so that the system operation efficiency is improved; and the two models are retrained periodically by using the preferred data output selector according to the output result of the models, so that the self-adaptive model adjustment of different transformers is realized, and the continuous learning and self-improvement of the learning model are realized.
Description
Technical Field
The invention belongs to the technical field of transformer fault diagnosis, and particularly relates to a transformer fault diagnosis device and method based on logistic regression and naive Bayes.
Background
The power transformer is one of the vital devices in the power system, and the operation condition of the power transformer is directly related to the safe and stable operation of the power system. The transformer state evaluation and the fault detection of the branch subject based on the time sequence data are always the subject of the key research in the industry, and have important significance for the stable operation of the power grid.
With the development of energy internet, the traditional evaluation scheme and the single machine learning transformer evaluation scheme are initially explored, and although the methods have certain effects, the problems that the fault judgment accuracy is low, the external environment change cannot be updated in time and the like exist. In recent years, scholars have proposed more hierarchical machine learning algorithms to upgrade diagnostic methods, but still have some disadvantages, such as: the fuzzy clustering method has no ideal effect on large-scale sample classification; SVM is also difficult to apply to large-scale training samples, etc.
Disclosure of Invention
Based on the above problems in the prior art, the present invention is directed to a transformer fault diagnosis apparatus and method based on logistic regression and naive bayes.
In order to achieve the above purpose, the invention adopts the technical scheme that:
a transformer fault diagnosis device based on logistic regression and naive Bayes comprises:
the serial data input interface is used for adapting to a communication interface of the power system and integrating historical state data and historical environment data;
the logistic regression continuous machine learning model implementation module is used for calling an internally packaged logistic regression algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating and training the machine learning model and generating a transformer fault identification data prediction result;
the naive Bayes continuous machine learning model implementation module is used for calling an internally packaged naive Bayes algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating the training machine learning model and generating a transformer fault identification data prediction result;
the preferred data output selector is used for receiving the transformer fault identification data prediction result output by the logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module, evaluating the prediction hit rate of the prediction result, selecting a model with a higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameter with a lower prediction hit rate;
and the prediction data output module is used for outputting the prediction result selected by the preferred data output selector.
Further, the series of data input interfaces are specifically used for inputting dependent variable data and multi-column independent variable data, the dependent variable data comprise state factors, and the independent variable data comprise environment data, oil chromatography collection data and operation state data.
Further, the preferential data output selector is specifically configured to evaluate a prediction hit rate of a prediction result of the transformer fault identification data output by the logistic regression continuous machine learning model implementation module and the naive bayes continuous machine learning model implementation module in a set period, and select a model with a higher prediction hit rate as a diagnostic model; and after the prediction results of the transformer fault identification data in the next period are received, the prediction hit rates of the two models are reevaluated, and the model priority is dynamically adjusted.
Further, the prediction data output module specifically includes a display unit and/or an HTTP interface, the display unit is configured to display, to a user, a transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector, and the HTTP interface is configured to transmit the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to an external system.
And the standard embedded computing support hardware module is used for providing external computing resources required for supporting the operation of the logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module.
Based on the transformer fault diagnosis device, the invention also provides a transformer fault diagnosis method based on logistic regression and naive Bayes, which comprises the following steps:
s1, adapting a communication interface of the power system by using the serial data input interface, and integrating historical state data and historical environment data;
s2, calling an internally packaged logistic regression algorithm by using a logistic regression continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s3, calling an internally packaged naive Bayes algorithm by using a naive Bayes continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s4, receiving the transformer fault identification data prediction results generated in the steps S2 and S3 by using a preferred data output selector, evaluating the prediction hit rate of the prediction results, selecting a model with higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameters with lower prediction hit rate;
and S5, outputting the prediction result selected in the step S4 by using a prediction data output module.
Further, in the step S1, a series of data input interfaces are specifically used to input dependent variable data and multiple columns of independent variable data, where the dependent variable data includes a state factor, and the independent variable data includes environmental data, oil chromatography collected data, and operating state data.
Further, in step S4, a preferred data output selector is specifically used to evaluate the prediction hit rate of the prediction result of the transformer fault identification data output by the logistic regression persistent machine learning model implementation module and the naive bayes persistent machine learning model implementation module in a set period, and a model with a higher prediction hit rate is selected as a diagnostic model; and after the prediction results of the transformer fault identification data in the next period are received, the prediction hit rates of the two models are reevaluated, and the model priority is dynamically adjusted.
Further, the predicted data output module in step S5 specifically includes a display unit and/or an HTTP interface, where the display unit is configured to show the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to a user, and the HTTP interface is configured to transmit the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to an external system.
Further, the method also comprises the step of providing external computing resources required for supporting the operation of the logistic regression continuous machine learning model implementation module in the step S3 and the naive Bayes continuous machine learning model implementation module in the step S4 by using a standard embedded computing support hardware module.
The invention has the beneficial effects that:
(1) the logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module provided by the invention encapsulate the training result model encapsulated by the algorithm in the module in a hardware form, and a multi-core CPU is selected as a hardware support, so that the system operation efficiency is improved;
(2) the preferred data output selector provided by the invention realizes that the two models are retrained periodically according to the output result of the model and the naive Bayes continuous machine learning model by the logistic regression continuous machine learning model, thereby realizing the self-adaptive model adjustment of different transformers and realizing the continuous learning and self-improvement of the learning model.
Drawings
FIG. 1 is a schematic diagram showing the results of a transformer fault diagnosis device based on logistic regression and naive Bayes provided by the invention;
fig. 2 is a schematic flow chart of a transformer fault diagnosis method based on logistic regression and naive bayes provided by the invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, the present embodiment provides a transformer fault diagnosis apparatus based on logistic regression and naive bayes, including:
the serial data input interface is used for adapting to a communication interface of the power system and integrating historical state data and historical environment data;
the logistic regression continuous machine learning model implementation module is used for calling an internally packaged logistic regression algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating and training the machine learning model and generating a transformer fault identification data prediction result;
the naive Bayes continuous machine learning model implementation module is used for calling an internally packaged naive Bayes algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating the training machine learning model and generating a transformer fault identification data prediction result;
the preferred data output selector is used for receiving the transformer fault identification data prediction result output by the logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module, evaluating the prediction hit rate of the prediction result, selecting a model with a higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameter with a lower prediction hit rate;
and the prediction data output module is used for outputting the prediction result selected by the preferred data output selector.
In the embodiment, the serial data input interface is used for adapting a communication interface of the power system, inputting dependent variable data and multi-column independent variable data, wherein the dependent variable data comprises a state factor, the independent variable data comprises environmental data, oil chromatogram acquisition data and running state data, and integrating the input dependent variable data and the multi-column independent variable data into stomach historical state data and historical environmental data. The function can be realized by adopting multiple groups of standard physical communication interfaces, such as a USB interface, a network port, a Type-C interface, WIFI and the like, and supporting protocols such as TCP/IP and serial communication and the like, and meanwhile, a standard API for data input is provided, and the standard data input format can be realized.
In this embodiment, the present invention constructs a logistic regression continuous machine learning model implementation module and a naive bayes continuous machine learning model implementation module respectively according to the idea of fully exerting the self-updating capability of the machine learning algorithm and hardwarzing the machine learning algorithm.
Logistic regression predicts the probability of an event occurring by fitting a logistic function/hypothesis function; naive Bayes Classification (NBC) is a method based on Bayes' theorem and assuming mutual independence between feature conditions, learning a joint probability distribution from input to output by a given training set and assuming independence between feature words as a premise, and then calculating an output which maximizes the posterior probability after input based on the learned model. The naive Bayes algorithm assumes that the attributes of the data sets are mutually independent, so the logic of the algorithm is very simple, the algorithm is stable, and when the data presents different characteristics, the classification performance of the naive Bayes is not different greatly. In other words, the naive bayes algorithm is more robust and does not exhibit too much difference for different types of data sets. When the relationship between the data set attributes is relatively independent, the naive Bayes classification algorithm has a better effect.
The logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module are respectively used for calling an internally packaged logistic regression algorithm and a naive Bayes algorithm to process integrated historical state data and historical environment data input by the series data input interface module, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result, namely whether a fault or sub-health possibly occurs. The function can be realized by adopting a piece of machine learning optimization hardware and codes, such as a GPU and a programmable development board. After the model code is determined, the programmable development board may be cured to standard hardware or firmware to implement a hardware implementation of the algorithm. The invention encapsulates the training result model encapsulated by the algorithm in a module in a hardware form, and selects the multi-core CPU as a hardware support, thereby improving the system operation efficiency.
In this embodiment, in order to solve the problem that the conventional oil chromatography and common machine learning model evaluation scheme does not have a multi-model preferred selection function and an automatic lag model parameter adjustment function, the present invention uses a preferred data output selector to perform statistical regression on the prediction hit rate of the transformer fault identification data prediction result output by a continuous machine learning model implementation module and a naive bayes continuous machine learning model implementation module in a set period for evaluation, selects a model with a higher prediction hit rate as a diagnosis model, and returns the diagnosis result to a prediction data output module, thereby improving the prediction hit rate of the final output result; after receiving the prediction result of the transformer fault identification data of the next period, re-evaluating the prediction hit rates of the two models, and dynamically adjusting the model priority; the preferred data output selector automatically adjusts relevant parameters of the model, such as the threshold value of AIC in stepwise logistic regression, the distribution function of data in the naive Bayes model and the like, when the predicted hit rate of one of the models lags against the selected logistic regression algorithm and the naive Bayes algorithm.
The preferred data output selector periodically retrains the two models according to the output result of the logistic regression continuous machine learning model realization module and the naive Bayes continuous machine learning model, thereby realizing the self-adaptive model adjustment of different transformers and realizing the continuous learning and self-improvement of the learning model.
The function of the preferred data output selector can be realized by adopting an embedded programmable development board, and the key point of the preferred data output selector is that the preferred data output selector can receive output data of a logistic regression continuous machine learning model realization module and a naive Bayes continuous machine learning model, evaluate historical performances and preferentially output result data.
In this embodiment, the prediction data output module specifically includes a display unit and/or an HTTP interface, where the display unit is configured to show, to a user, a transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector, and the HTTP interface is configured to transmit the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to an external system.
The above functions of the prediction data output module can be realized by adopting a display and codes, the working environment of the prediction data output module can be Python or Java and the like, and the existing open source chart system such as G2 and the like is called for drawing, so that the corresponding functions can be realized quickly and conveniently.
In this embodiment, in order to improve the computation efficiency of the logistic regression persistent machine learning model implementation module and the naive bayes persistent machine learning model implementation module, the present invention further includes a standard embedded computation support hardware module, which is used to provide external computation resources, such as result caching, log recording, setting parameter reading and writing, etc., required for supporting the operation of the logistic regression persistent machine learning model implementation module and the naive bayes persistent machine learning model implementation module. .
The functions of the standard embedded computing support hardware module can be realized by adopting a composite mainboard, and a storage module, a bus and the like are embedded in the mainboard, so that the computing resources of a logistic regression continuous machine learning model realization module and a naive Bayes continuous machine learning model realization module are provided and supported.
According to the invention, through mutual cooperation of all modules, according to the idea of fully exerting machine learning capacity and realizing hardware, historical state data, historical environment data and the like after the transformer is integrated are input into the algorithm module, and result data is preferentially output according to a recent model expression result, so that the recent fault prediction of the transformer is realized.
The method takes learning model hardware as a thought, a cyclic process of training the model and predicting future data is packaged into a system, and simultaneously, logistic regression and naive Bayes suitable for a prediction system are selected as a collocation algorithm, so that an innovative transformer equipment state evaluation scheme is realized, and further, the upgrading from basic tabular evaluation to intelligent continuous learning evaluation is realized, thereby contributing an innovative force to the upgrading of the production, operation and maintenance capacity of the power system in China.
Example 2
Based on the transformer fault diagnosis device, the invention also provides a transformer fault diagnosis method based on logistic regression and naive Bayes, as shown in FIG. 2, comprising the following steps:
s1, adapting a communication interface of the power system by using the serial data input interface, and integrating historical state data and historical environment data;
s2, calling an internally packaged logistic regression algorithm by using a logistic regression continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s3, calling an internally packaged naive Bayes algorithm by using a naive Bayes continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s4, receiving the transformer fault identification data prediction results generated in the steps S2 and S3 by using a preferred data output selector, evaluating the prediction hit rate of the prediction results, selecting a model with higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameters with lower prediction hit rate;
and S5, outputting the prediction result selected in the step S4 by using a prediction data output module.
In step S1, the present invention adapts a communication interface of a power system using a serial data input interface, inputs dependent variable data and multi-column independent variable data, wherein the dependent variable data includes a state factor, and the independent variable data includes environmental data, oil chromatogram collection data, and operating state data, and integrates the input dependent variable data and the multi-column independent variable data into stomach historical state data and historical environmental data.
In step S2 and step S3, the present invention constructs a logistic regression persistent machine learning model implementation module and a naive bayes persistent machine learning model implementation module, respectively, according to the idea of fully utilizing the self-updating capability of the machine learning algorithm and hardwarzing the same.
The logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module are respectively used for calling an internally packaged logistic regression algorithm and an naive Bayes algorithm to process integrated historical state data and historical environment data input by the series data input interface module, automatically updating the training machine learning model and generating a transformer fault identification data prediction result, namely whether a fault or sub-health and the like possibly occurs
The invention encapsulates the training result model encapsulated by the algorithm in a module in a hardware form, and selects the multi-core CPU as a hardware support, thereby improving the system operation efficiency.
In step S4, in order to solve the problem that the conventional oil chromatography and common machine learning model evaluation schemes do not have a multi-model preferred selection function and an automatic parameter adjustment function of a laggard model, the present invention uses a preferred data output selector to perform statistical regression on a continuous machine learning model implementation module and a naive bayes continuous machine learning model implementation module to perform evaluation on the prediction hit rate of the transformer fault identification data prediction result output by the module in a set period, selects a model with a higher prediction hit rate as a diagnosis model, and returns the diagnosis result to a prediction data output module, thereby improving the prediction hit rate of the final output result; after receiving the prediction result of the transformer fault identification data of the next period, re-evaluating the prediction hit rates of the two models, and dynamically adjusting the model priority; the preferred data output selector automatically adjusts relevant parameters of the model, such as the threshold value of AIC in stepwise logistic regression, the distribution function of data in the naive Bayes model and the like, when the predicted hit rate of one of the models lags against the selected logistic regression algorithm and the naive Bayes algorithm.
The preferred data output selector periodically retrains the two models according to the output result of the logistic regression continuous machine learning model realization module and the naive Bayes continuous machine learning model, thereby realizing the self-adaptive model adjustment of different transformers and realizing the continuous learning and self-improvement of the learning model.
In step S5, the prediction data output module of the present invention specifically includes a display unit and/or an HTTP interface, where the display unit is configured to display the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to a user, and the HTTP interface is configured to transmit the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to an external system.
In addition, in order to improve the calculation efficiency of the logistic regression model implementation module and the naive Bayes model implementation module, the invention also comprises an external calculation resource which is provided by a standard embedded calculation support hardware module and is required for supporting the logistic regression model implementation module and the naive Bayes model implementation module to operate, such as result caching, log recording, parameter setting reading and writing and the like. .
The above examples of the present invention are merely illustrative of the present invention and are not intended to limit the embodiments of the present invention. Variations and modifications in other variations will occur to those skilled in the art upon reading the foregoing description. Not all embodiments are exhaustive. All obvious changes and modifications of the present invention are within the scope of the present invention.
Claims (10)
1. A transformer fault diagnosis device based on logistic regression and naive Bayes is characterized by comprising the following steps:
the serial data input interface is used for adapting to a communication interface of the power system and integrating historical state data and historical environment data;
the logistic regression continuous machine learning model implementation module is used for calling an internally packaged logistic regression algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating and training the machine learning model and generating a transformer fault identification data prediction result;
the naive Bayes continuous machine learning model implementation module is used for calling an internally packaged naive Bayes algorithm to process the integrated historical state data and historical environment data input by the series of data input interface modules, automatically updating the training machine learning model and generating a transformer fault identification data prediction result;
the preferred data output selector is used for receiving the transformer fault identification data prediction result output by the logistic regression continuous machine learning model implementation module and the naive Bayes continuous machine learning model implementation module, evaluating the prediction hit rate of the prediction result, selecting a model with a higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameter with a lower prediction hit rate;
and the prediction data output module is used for outputting the prediction result selected by the preferred data output selector.
2. The transformer fault diagnosis device based on logistic regression and naive Bayes as claimed in claim 1, wherein said series of data input interfaces are specifically used for inputting dependent variable data and multi-column independent variable data, said dependent variable data comprises state factors, and said independent variable data comprises environment data, oil chromatogram collected data and operation state data.
3. The transformer fault diagnosis device based on logistic regression and naive Bayes as claimed in claim 1, wherein said preferential data output selector is specifically configured to evaluate a prediction hit rate of a transformer fault identification data prediction result output by said logistic regression persistent machine learning model implementation module and naive Bayes persistent machine learning model implementation module in a set period, and select a model with a higher prediction hit rate as a diagnosis model; and after the prediction results of the transformer fault identification data in the next period are received, the prediction hit rates of the two models are reevaluated, and the model priority is dynamically adjusted.
4. The transformer fault diagnosis device based on logistic regression and naive bayes as claimed in claim 3, wherein said prediction data output module specifically comprises a display unit and/or an HTTP interface, said display unit is used for displaying the transformer fault diagnosis result outputted by the diagnosis model selected by said preferential data output selector to the user, said HTTP interface is used for transmitting the transformer fault diagnosis result outputted by the diagnosis model selected by said preferential data output selector to the external system.
5. The transformer fault diagnosis device based on the logistic regression and naive bayes according to any of claims 1 to 4, further comprising a standard embedded computing support hardware module for providing external computing resources required for supporting the operation of the logistic regression persistent machine learning model implementation module and the naive bayes persistent machine learning model implementation module.
6. A transformer fault diagnosis method based on logistic regression and naive Bayes is characterized by comprising the following steps:
s1, adapting a communication interface of the power system by using the serial data input interface, and integrating historical state data and historical environment data;
s2, calling an internally packaged logistic regression algorithm by using a logistic regression continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s3, calling an internally packaged naive Bayes algorithm by using a naive Bayes continuous machine learning model implementation module to process the historical state data and the historical environment data integrated in the step S1, automatically updating the training machine learning model, and generating a transformer fault identification data prediction result;
s4, receiving the transformer fault identification data prediction results generated in the steps S2 and S3 by using a preferred data output selector, evaluating the prediction hit rate of the prediction results, selecting a model with higher prediction hit rate as a diagnosis model, and automatically adjusting the model parameters with lower prediction hit rate;
and S5, outputting the prediction result selected in the step S4 by using a prediction data output module.
7. The transformer fault diagnosis method based on logistic regression and naive Bayes as claimed in claim 6, wherein said step S1 inputs dependent variable data and multi-column independent variable data specifically using series data input interface, said dependent variable data comprising state factor, said independent variable data comprising environment data, oil chromatogram collected data and operation state data.
8. The transformer fault diagnosis method based on logistic regression and naive Bayes as claimed in claim 6, wherein said step S4 specifically utilizes a preferred data output selector to evaluate the prediction hit rate of the transformer fault identification data prediction result outputted by said logistic regression persistent machine learning model implementation module and naive Bayes persistent machine learning model implementation module in a set period, and selects a model with higher prediction hit rate as a diagnosis model; and after the prediction results of the transformer fault identification data in the next period are received, the prediction hit rates of the two models are reevaluated, and the model priority is dynamically adjusted.
9. The transformer fault diagnosis device according to claim 8, wherein the prediction data output module in step S5 specifically includes a display unit and/or an HTTP interface, the display unit is configured to display the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to a user, and the HTTP interface is configured to transmit the transformer fault diagnosis result output by the diagnosis model selected by the preferential data output selector to an external system.
10. The transformer fault diagnosis device based on the logistic regression and naive bayes according to any of claims 6 to 9, further comprising providing external computing resources required for supporting the operation of the logistic regression persistent machine learning model implementation module in step S3 and the naive bayes persistent machine learning model implementation module in step S4 by using a standard embedded computing support hardware module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010472929.XA CN111723519B (en) | 2020-05-29 | 2020-05-29 | Transformer fault diagnosis device and method based on logistic regression and naive Bayes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010472929.XA CN111723519B (en) | 2020-05-29 | 2020-05-29 | Transformer fault diagnosis device and method based on logistic regression and naive Bayes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111723519A true CN111723519A (en) | 2020-09-29 |
| CN111723519B CN111723519B (en) | 2022-11-15 |
Family
ID=72565277
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010472929.XA Active CN111723519B (en) | 2020-05-29 | 2020-05-29 | Transformer fault diagnosis device and method based on logistic regression and naive Bayes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111723519B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112651444A (en) * | 2020-12-29 | 2021-04-13 | 山东科技大学 | Self-learning-based non-stationary process anomaly detection method |
| CN112990592A (en) * | 2021-03-26 | 2021-06-18 | 广东工业大学 | Shared vehicle fault prediction method and system |
| CN115456210A (en) * | 2022-08-22 | 2022-12-09 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power utilization complaint early warning method based on cascade logistic regression Bayesian algorithm |
| WO2023045512A1 (en) * | 2021-09-24 | 2023-03-30 | 浪潮集团有限公司 | Method and system for performing fault detection on industrial hardware on basis of machine learning |
| CN117057486A (en) * | 2023-10-11 | 2023-11-14 | 云南电投绿能科技有限公司 | Operation and maintenance cost prediction method, device and equipment for power system and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100145647A1 (en) * | 2008-12-04 | 2010-06-10 | Xerox Corporation | System and method for improving failure detection using collective intelligence with end-user feedback |
| CN108051660A (en) * | 2017-10-31 | 2018-05-18 | 华北电力大学（保定） | A kind of transformer fault combined diagnosis method for establishing model and diagnostic method |
| CN110162014A (en) * | 2019-05-29 | 2019-08-23 | 上海理工大学 | A kind of breakdown of refrigeration system diagnostic method of integrated multi-intelligence algorithm |
| CN110222723A (en) * | 2019-05-14 | 2019-09-10 | 华南理工大学 | A kind of starting prediction technique of football match based on mixed model |
| CN110968069A (en) * | 2018-09-28 | 2020-04-07 | 新疆金风科技股份有限公司 | Fault prediction method of wind generating set, corresponding device and electronic equipment |
| CN111105063A (en) * | 2018-10-26 | 2020-05-05 | 北京国双科技有限公司 | Fault prediction method, fault prediction device, model construction method, fault prediction device, processor and readable storage medium |
-
2020
- 2020-05-29 CN CN202010472929.XA patent/CN111723519B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100145647A1 (en) * | 2008-12-04 | 2010-06-10 | Xerox Corporation | System and method for improving failure detection using collective intelligence with end-user feedback |
| CN108051660A (en) * | 2017-10-31 | 2018-05-18 | 华北电力大学（保定） | A kind of transformer fault combined diagnosis method for establishing model and diagnostic method |
| CN110968069A (en) * | 2018-09-28 | 2020-04-07 | 新疆金风科技股份有限公司 | Fault prediction method of wind generating set, corresponding device and electronic equipment |
| CN111105063A (en) * | 2018-10-26 | 2020-05-05 | 北京国双科技有限公司 | Fault prediction method, fault prediction device, model construction method, fault prediction device, processor and readable storage medium |
| CN110222723A (en) * | 2019-05-14 | 2019-09-10 | 华南理工大学 | A kind of starting prediction technique of football match based on mixed model |
| CN110162014A (en) * | 2019-05-29 | 2019-08-23 | 上海理工大学 | A kind of breakdown of refrigeration system diagnostic method of integrated multi-intelligence algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 刘岩 等: "基于机器学习决策树的计量设备异常分析", 《自动化与仪器仪表》, 25 May 2018 (2018-05-25), pages 171 - 174 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112651444A (en) * | 2020-12-29 | 2021-04-13 | 山东科技大学 | Self-learning-based non-stationary process anomaly detection method |
| CN112651444B (en) * | 2020-12-29 | 2022-08-02 | 山东科技大学 | Self-learning-based non-stationary process anomaly detection method |
| CN112990592A (en) * | 2021-03-26 | 2021-06-18 | 广东工业大学 | Shared vehicle fault prediction method and system |
| WO2023045512A1 (en) * | 2021-09-24 | 2023-03-30 | 浪潮集团有限公司 | Method and system for performing fault detection on industrial hardware on basis of machine learning |
| CN115456210A (en) * | 2022-08-22 | 2022-12-09 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power utilization complaint early warning method based on cascade logistic regression Bayesian algorithm |
| CN115456210B (en) * | 2022-08-22 | 2024-04-12 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power consumption complaint early warning method based on cascading logistic regression Bayesian algorithm |
| CN117057486A (en) * | 2023-10-11 | 2023-11-14 | 云南电投绿能科技有限公司 | Operation and maintenance cost prediction method, device and equipment for power system and storage medium |
| CN117057486B (en) * | 2023-10-11 | 2023-12-22 | 云南电投绿能科技有限公司 | Operation and maintenance cost prediction method, device and equipment for power system and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111723519B (en) | 2022-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111723519B (en) | Transformer fault diagnosis device and method based on logistic regression and naive Bayes | |
| CN116846987B (en) | Interactive interface generation method and system of industrial Internet | |
| CN111723520A (en) | Transformer fault diagnosis device and method based on decision tree and random forest | |
| CN102955977A (en) | Energy efficiency service method and energy efficiency service platform adopting same on basis of cloud technology | |
| CN111091240A (en) | Public institution electric power energy efficiency monitoring system and service method | |
| CN116760772B (en) | Control system and method for converging flow divider | |
| CN107480432B (en) | Load decomposition method based on cloud platform | |
| WO2024065988A1 (en) | Device monitoring method and system, electronic device, and storage medium | |
| CN108009089B (en) | Incremental machine learning method and system based on transparent computing | |
| CN116896495A (en) | Bian Yun collaborative model dynamic training method | |
| CN117666448B (en) | Energy saving method, device, equipment and medium based on Internet of things energy consumption prediction | |
| Tripathi et al. | Data-driven optimizations in IoT: A new frontier of challenges and opportunities | |
| CN118101361A (en) | Data acquisition terminal and method for air conditioner monitoring system | |
| CN115576203A (en) | Embedded electric precipitation intelligent control method and system based on neural network | |
| CN111723518A (en) | Transformer fault diagnosis device and method based on condition inference tree and AdaBoost | |
| CN201256296Y (en) | Intelligent integrated system for reactive compensation and reactive optimization for electricity distribution network | |
| CN109001985B (en) | Photovoltaic power station modeling method and device based on codes | |
| US11924054B1 (en) | Managing remote terminal communications | |
| CN111597860A (en) | System and method for CNC (computer numerical control) cutter breakage prediction | |
| CN110659681A (en) | Time sequence data prediction system and method based on pattern recognition | |
| CN117171551B (en) | Large-scale industrial equipment data analysis and intelligent management method | |
| Chen et al. | Research on Data Collection and Observability of Gas Station in Cloud Environment | |
| CN115759236B (en) | Model training method, information sending method, device, equipment and medium | |
| US20240039820A1 (en) | Messaging protocol for configuring remote terminal unit | |
| KR102334975B1 (en) | System For Controlling An Automate Equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |