CN110135064B - Method, system and controller for predicting temperature faults of rear bearing of generator - Google Patents
Method, system and controller for predicting temperature faults of rear bearing of generator Download PDFInfo
- Publication number
- CN110135064B CN110135064B CN201910402698.2A CN201910402698A CN110135064B CN 110135064 B CN110135064 B CN 110135064B CN 201910402698 A CN201910402698 A CN 201910402698A CN 110135064 B CN110135064 B CN 110135064B
- Authority
- CN
- China
- Prior art keywords
- generator
- temperature
- bearing
- parameter
- rear bearing
- 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.)
- Active
Links
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Control Of Eletrric Generators (AREA)
Abstract
The invention provides a method, a system and a controller for predicting temperature faults of a rear bearing of a generator, wherein the method comprises the following steps: acquiring data of fan operation state parameters, wherein the data comprises historical data and current data; acquiring state parameters related to the temperature of a rear bearing of the generator according to the data; constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model by utilizing the state parameters; and predicting the temperature of the rear bearing of the generator and/or the residual time of the temperature alarm fault of the rear bearing of the generator. The invention can accurately predict the temperature of the bearing behind the generator and the residual time from the occurrence of the alarm fault of the temperature of the bearing behind the generator, thereby providing guiding advice and help for site engineers to take effective and reasonable protective measures.
Description
Technical Field
The invention relates to the technical field of fault prediction of mechanical equipment, in particular to the technical field of temperature fault prediction of a rear bearing of a fan generator, and particularly relates to a method, a system and a controller for predicting the temperature fault of the rear bearing of the generator.
Background
The generator is one of key parts of the fan, the fault maintenance cost of the generator system accounts for about 10% of the healthy operation and maintenance cost of the whole wind turbine generator, the related faults of the generator are one of the most main causes for the shutdown of the wind turbine generator, and the average fault removal time is long. Therefore, the healthy operation and maintenance of the generator is a very important part in the intelligent operation and maintenance of the wind turbine. In the process of on-line state monitoring of the generator, a field engineer focuses not only on the current running state of the generator system, but also on the problem of fault prediction of key faults in the generator. The temperature fault of the bearing behind the generator can be found out to be a more frequent fault of the generator by counting the main faults related to the generator in the historical fault record of the fan. Once the temperature alarm fault of the bearing after the generator occurs, the generator is seriously influenced, the fan is stopped due to the fault, and the generated energy of the whole fan is finally influenced. Therefore, the future bearing temperature fault after the generator is predicted, the future trend of the bearing temperature after the generator and the emergency degree of the bearing temperature alarm fault after the generator can be known in time, and guiding suggestions and assistance are provided for site engineers to take effective and reasonable protective measures. However, the prediction accuracy of the temperature fault of the bearing behind the fan is not high at present, so that the prediction accuracy has little significance for guiding actual work.
Disclosure of Invention
In view of the above drawbacks of the prior art, the present invention is directed to a method, a system and a controller for predicting a temperature failure of a rear bearing of a generator, which are used for solving the problem that in the prior art, the accuracy of predicting the temperature failure of the rear bearing of a fan is not high.
To achieve the above and other related objects, the present invention provides a method for predicting a temperature failure of a rear bearing of a generator, including: acquiring data of fan operation state parameters, wherein the data comprises historical data and current data; acquiring state parameters related to the temperature of a rear bearing of the generator according to the data; constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model by utilizing the state parameters; and predicting the temperature of the rear bearing of the generator and/or the residual time of the temperature alarm fault of the rear bearing of the generator.
In an embodiment of the present invention, the implementation manner of obtaining the state parameter related to the temperature of the bearing behind the generator according to the data includes: analyzing the data by using a correlation analysis method and obtaining a correlation analysis result; and selecting a state parameter with the correlation larger than a correlation threshold value in the correlation analysis result as a state parameter related to the temperature of the rear bearing of the generator.
In an embodiment of the present invention, the correlation coefficients used in the correlation analysis method include Pearson correlation coefficients and Maximum Information Coefficients (MIC).
In an embodiment of the present invention, the implementation manner of constructing the post-generator bearing temperature prediction model based on the SVR model and/or the post-generator bearing temperature alarm fault remaining time prediction model based on the SVR model by using the state parameters includes: acquiring parameter weights of the state parameters related to the temperature of the rear bearing of the generator in the prediction model; obtaining optimal super parameters by adopting a mode of combining grid search and cross verification; taking the temperature of the generator rear bearing at the continuous N moments and the state parameters related to the temperature of the generator rear bearing at the continuous N moments in the time sequence as the input of an SVR model, taking the temperature of the generator rear bearing at the future T moment as the output, and constructing the generator rear bearing temperature prediction model based on the SVR model by combining the parameter weight and the optimal super parameter; carrying out smoothing treatment on the state parameter data related to the temperature of the rear bearing of the generator to obtain smoothed state parameter related to the temperature of the rear bearing of the generator; and taking the state parameters related to the temperature of the rear bearing of the generator after the smoothing treatment as the input of an SVR model, taking the residual time of the distance fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct a residual time prediction model of the temperature alarm fault of the rear bearing of the generator based on the SVR model.
In an embodiment of the present invention, the implementation manner of obtaining the parameter weight of the state parameter related to the post-generator bearing temperature in the prediction model is: for the ith state parameter, the parameter weight coefficientThe method comprises the following steps:
w i ＝r i ×MIC i
wherein w is i For a calculated intermediate value, r i Pearson correlation coefficient, MIC, representing the state parameter and the temperature of the rear bearing of the generator i MIC correlation value representing the state parameter and the temperature of the bearing behind the generator, M representing the number of relevant state parameters.
In an embodiment of the present invention, the smoothing method is: considering the current moment in the time sequenceThe data point and the previous h-1 data point are used for relating the temperature fault related state parameter P of the bearing after the generator at the moment t i(t) The method comprises the following steps:
where the meaning of h is to consider h data points in the time series.
In an embodiment of the present invention, there is also provided a post-generator bearing temperature failure prediction system, including: the data acquisition module is used for acquiring data of the running state parameters of the fan, wherein the data comprise historical data and current data; the related state parameter acquisition module is connected with the data acquisition module and is used for acquiring state parameters related to the temperature of the rear bearing of the generator according to the data; the model construction module is connected with the related state parameter acquisition module and is used for constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model; and the prediction module is respectively connected with the model construction module and the data acquisition module and is used for predicting the temperature of the bearing behind the generator and the residual time of the bearing temperature alarm fault after the generator.
In an embodiment of the invention, the model building module includes: the parameter weight acquisition unit is used for acquiring the parameter weight of the state parameter related to the temperature of the rear bearing of the generator in the prediction model; the super-parameter acquisition unit is used for acquiring optimal super-parameters; the smoothing processing unit is used for obtaining the temperature-related state parameters of the rear bearing of the generator after the smoothing processing; the temperature prediction unit is respectively connected with the parameter weight acquisition unit and the super parameter acquisition unit and is used for constructing the SVR model-based generator rear bearing temperature prediction model; the fault remaining time prediction unit is respectively connected with the parameter weight acquisition unit, the super parameter acquisition unit and the smoothing processing unit and is used for constructing the SVR model-based bearing temperature alarm fault remaining time prediction model after the generator.
In an embodiment of the present invention, there is also provided a computer-readable storage medium having stored thereon a computer program, characterized in that the program when called by a processor implements the post-generator bearing temperature failure prediction method as described above.
In an embodiment of the present invention, there is further provided a controller, including: a memory for storing a computer program; and the processor is in communication connection with the memory and is used for realizing the method for predicting the temperature failure of the rear bearing of the generator when the computer program is called.
As described above, the method, the system and the controller for predicting the temperature fault of the rear bearing of the motor have the following beneficial effects: firstly, in the design process, the invention calculates the correlation of all state parameters in an SCADA (Supervisory Control And Data Acquisition, data acquisition and monitoring control) system, and selects the state parameters with stronger correlation with the temperature of a bearing behind a generator from the calculated state parameters, thereby ensuring the comprehensiveness of the input parameters of the prediction system; secondly, according to the invention, the state parameters related to the temperature of the bearing behind the generator are selected through a correlation analysis method, the influence of each state parameter on the temperature of the bearing behind the generator is considered, the weight of each state parameter is designed according to the correlation analysis result, the parameter data is smoothed according to a moving smoothing method, the optimal super-parameters are searched in a mode of combining grid search and cross verification, the time sequence of the temperature of the bearing behind the generator and the related state parameters is considered, the temperature of the bearing behind the generator and the related state parameters at N continuous moments in the time sequence are taken as the input of a model, and a SVR model-based prediction model of the temperature of the bearing behind the generator and/or a prediction model of the residual time of alarm fault of the temperature of the bearing behind the generator are established, so that the prediction precision is higher; thirdly, the invention not only considers the prediction of the temperature of the rear bearing of the generator, but also predicts the residual time of the alarm fault of the temperature of the rear bearing of the generator, thereby providing guiding advice and assistance for on-site engineers to take effective and reasonable protective measures.
Drawings
Fig. 1 is a schematic diagram of an implementation process of a method for predicting a temperature failure of a rear bearing of a generator according to an embodiment of the present invention.
Fig. 2 is a diagram showing an example of a post-generator bearing temperature prediction result obtained by the post-generator bearing temperature fault prediction method according to an embodiment of the present invention.
Fig. 3 is a diagram showing an example of a prediction result of the residual time of the post-generator bearing temperature alarm fault obtained by the post-generator bearing temperature fault prediction method according to the embodiment of the invention.
Fig. 4 is a schematic diagram of a process for constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on an SVR model by using the state parameters according to an embodiment of the present invention.
Fig. 5 is a schematic structural diagram of an implementation of a method for predicting a temperature failure of a rear bearing of a generator according to an embodiment of the present invention.
FIG. 6 is a schematic diagram of an implementation structure of a model building block according to an embodiment of the present invention.
Description of element reference numerals
200. Temperature fault prediction system for rear bearing of generator
210. Data acquisition module
220. Related state parameter acquisition module
230. Model building module
240. Prediction module
231. Parameter weight acquisition unit
232. Super parameter acquisition unit
233. Smoothing processing unit
234. Temperature prediction unit
235. Residual time of failure prediction unit
S110 to S140 steps
S131 to S135 steps
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the following embodiments and features in the embodiments may be combined with each other without conflict.
It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention by way of illustration, and only the components related to the present invention are shown in the illustrations, not according to the number, shape and size of the components in actual implementation, and the form, number and proportion of each component in actual implementation may be arbitrarily changed, and the layout of the components may be more complex.
The temperature fault of the bearing behind the generator can be found out to be a more frequent fault of the generator by counting the main faults related to the generator in the historical fault record of the fan. Once the temperature alarm fault of the bearing after the generator occurs, the generator is seriously influenced, the fan is stopped due to the fault, and the like, and finally the generated energy of the whole fan is reduced. Therefore, the future temperature faults of the bearing behind the generator are predicted, the development trend of the temperature of the bearing behind the generator and the emergency degree of the temperature alarm faults of the bearing behind the generator can be known in time, and guiding suggestions and assistance are provided for site engineers to take effective and reasonable protective measures. However, the prediction accuracy of the temperature fault of the bearing behind the fan is not high at present, so that the prediction accuracy has little significance for guiding actual work. To solve this problem, in one embodiment of the present invention, there is provided a method for predicting a post-generator bearing temperature failure, including: acquiring data of fan operation state parameters, wherein the data comprises historical data and current data; acquiring state parameters related to the temperature of a rear bearing of the generator according to the data; constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model by utilizing the state parameters; and predicting the temperature of the rear bearing of the generator and/or the residual time of the temperature alarm fault of the rear bearing of the generator.
Referring to fig. 1, a schematic diagram of an implementation process of a method for predicting a temperature failure of a rear bearing of a generator according to an embodiment of the present invention is shown, where the method includes:
s110, acquiring data of fan operation state parameters, wherein the data comprises historical data and current data. The SCADA system is a computer-based production process control and scheduling automation system, can monitor and control on-site operation equipment, and is most widely applied to power systems at present and is most mature in technical development. Thus, all parameters in the SCADA system regarding the operational status of the fans can be directly invoked in this step. In consideration of the post-generator bearing temperature and the timing of the relevant state parameters, the invention not only selects the current data of the state parameters, but also takes the historical data of the state parameters into consideration so as to enhance the accuracy of the prediction result.
S120, acquiring state parameters related to the temperature of the rear bearing of the generator according to the data. And (5) analyzing and processing the state parameters acquired in the step S110, and further obtaining the state parameters related to the temperature of the rear bearing of the generator. Wherein the number of state parameters related to the temperature of the rear bearing of the generator is denoted as M.
S130, constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault remaining time prediction model based on the SVR model by utilizing the state parameters. The SVR refers to a support vector regression machine (Support Vactor Regerssion), and belongs to an important application branch of the Support Vector Machine (SVM). An important idea of SVR is to find a regression plane to which all data of a set is closest. Based on the state parameters related to the temperature of the rear bearing of the generator obtained in the step S120, a SVR model is utilized to construct a rear bearing temperature prediction model of the generator based on the SVR model and/or a residual time prediction model of the temperature alarm fault of the rear bearing of the generator based on the SVR model.
Specifically, firstly, the data sets of the M state parameters obtained in the step S120 are divided into a test set and a training set according to a certain proportion, preferably, the proportion is 2:8, the test set data are used for testing the prediction precision, and the training set data are used for training a prediction model; and then training the prediction model, training the SVR model by using the data set, and selecting the optimal system parameters to improve the prediction accuracy.
S140, predicting the temperature of the rear bearing of the generator and/or the residual time of the temperature alarm fault of the rear bearing of the generator. When the system monitors the running state degradation of the generator and gives an alarm, the temperature of the rear bearing of the generator and/or the residual time of the alarm fault of the temperature of the bearing after the generator is generated are respectively predicted by utilizing the prediction model according to the temperature related state parameter data of the rear bearing of the generator of the online SCADA system.
Fig. 2 is a diagram showing an example of a post-generator bearing temperature prediction result obtained by the post-generator bearing temperature fault prediction method according to an embodiment of the present invention. The temperature prediction model of the rear bearing of the generator can accurately predict the temperature of the rear bearing of the generator and track the change trend of the temperature prediction model of the rear bearing of the generator.
Fig. 3 is a diagram showing an example of a prediction result of the residual time of the post-generator bearing temperature alarm fault obtained by the post-generator bearing temperature fault prediction method according to the embodiment of the invention. The residual time prediction model for the bearing temperature alarm fault after the generator can accurately track the change of the residual time for the bearing temperature alarm fault after the generator.
Tables 1-1 and 1-2 are a comparison of the proposed method (GC-WP_SVR) with the commonly used predictions of several methods, including SVR models (PSO-WP_SVR) using particle swarm optimization algorithms (Particle Swarm Optimization, PSO) for super-parametric optimization, conventional SVR models without state parameter weighting, and ELM (Extreme Learning Machine ) models. Three evaluation indexes are adopted to quantitatively evaluate the accuracy of the fault prediction model: root mean square error RMSE, mean absolute error MAE, mean squareSquare correlation coefficient R 2 The prediction effect of the model is evaluated, and the table shows that the prediction accuracy is improved by the method provided by the invention.
TABLE 1-1 comparison of bearing temperature prediction model results after generator
TABLE 1-2 comparison of bearing temperature failure residual time prediction model results after Generator Generation
In one embodiment of the present invention, the implementation of obtaining the state parameter related to the temperature of the rear bearing of the generator according to the data includes:
analyzing the data by using a correlation analysis method and obtaining a correlation analysis result, wherein the correlation coefficients adopted in the correlation analysis method comprise but are not limited to Pearson correlation coefficients and Maximum Information Coefficients (MIC);
and selecting a state parameter with the correlation larger than a correlation threshold value in the correlation analysis result as a state parameter related to the temperature of the rear bearing of the generator. The correlation threshold is selected according to actual conditions, and the higher the correlation threshold is, the fewer parameters are selected; the lower the correlation threshold, the more parameters are selected but the more computationally intensive. Preferably, the correlation threshold is 0.9.
Specifically, in this embodiment, the correlation analysis method is used to calculate the correlation between each state parameter and the temperature of the bearing behind the generator, and the state parameter having a strong relationship with the temperature of the bearing behind the generator is selected, and the specifically selected state parameter is shown in table 2.
Table 2 related state variable table
In one embodiment of the present invention, the implementation of step S130 includes: acquiring parameter weights of the state parameters related to the temperature of the rear bearing of the generator in the prediction model; obtaining optimal super parameters by adopting a mode of combining grid search and cross verification; taking the temperature of the generator rear bearing at the continuous N moments and the state parameters related to the temperature of the generator rear bearing at the continuous N moments in time sequence as the input of an SVR model, wherein N is a positive integer, the temperature of the generator rear bearing at the future T moment is taken as the output, and combining the parameter weight and the optimal super parameter to construct the SVR model-based generator rear bearing temperature prediction model; carrying out smoothing treatment on the state parameter data related to the temperature of the rear bearing of the generator to obtain smoothed state parameter related to the temperature of the rear bearing of the generator; and taking the state parameters related to the temperature of the rear bearing of the generator after the smoothing treatment as the input of an SVR model, taking the residual time of the distance fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct a residual time prediction model of the temperature alarm fault of the rear bearing of the generator based on the SVR model. Referring to fig. 4, a schematic implementation process of constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault remaining time prediction model based on an SVR model by using the state parameters according to an embodiment of the present invention is shown, where the implementation process includes:
s131, acquiring parameter weights of the state parameters related to the temperature of the rear bearing of the generator in the prediction model;
s132, obtaining optimal super parameters by adopting a mode of combining grid search and cross verification;
s133, taking the temperature of the rear bearing of the generator at N continuous moments and the state parameters related to the temperature of the rear bearing of the generator at N continuous moments in time sequence as the input of an SVR model, wherein N is a positive integer, the temperature of the rear bearing of the generator at the time T future is taken as the output, and combining the parameter weight and the optimal super parameter to construct the SVR model-based temperature prediction model of the rear bearing of the generator;
s134, carrying out smoothing treatment on the state parameter data related to the temperature of the rear bearing of the generator to obtain smoothed state parameter related to the temperature of the rear bearing of the generator;
s135, taking the state parameters related to the temperature of the rear bearing of the generator after the smoothing treatment as the input of an SVR model, taking the residual time of the distance fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct a residual time prediction model of the temperature alarm fault of the rear bearing of the generator based on the SVR model.
The steps S131 to S135 are described in detail below.
S131, acquiring the parameter weight of the state parameter related to the temperature of the rear bearing of the generator in the prediction model. The weight is designed as follows:
w i ＝r i ×MIC i
wherein, the liquid crystal display device comprises a liquid crystal display device,weight coefficient for corresponding different state parameters, w i For a calculated intermediate value, r i Pearson correlation coefficient, MIC, representing the state parameter and the temperature of the rear bearing of the generator i MIC correlation value representing the state parameter and the temperature of the bearing behind the generator, M representing the number of relevant state parameters. Specifically, in the present embodiment, the obtained correlation results of the respective state parameters and the obtained weights are shown in table 3.
S132, obtaining the optimal super-parameters by adopting a mode of combining grid search and cross verification. The prediction model is an SVR model, an RBF kernel function is adopted as a kernel function of the SVR model, and the optimal super parameters to be searched are a penalty factor C and a kernel function g.
S133, taking the temperature of the rear bearing of the generator at N continuous moments and the state parameters related to the temperature of the rear bearing of the generator at N continuous moments in time sequence as the input of the SVR model, wherein N is a positive integer, the temperature of the rear bearing of the generator at the T moment in the future is taken as the output, and combining the parameter weight and the optimal super parameter to construct the SVR model-based temperature prediction model of the rear bearing of the generator. The SVR model-based generator rear bearing temperature prediction model specifically comprises: in the generator rear bearing temperature trend prediction model, the generator rear bearing temperature and related state parameters at N continuous moments in time sequence are taken as the input of the model, and the generator rear bearing temperature at the future moment T is taken as the output, taking the time sequence of the generator rear bearing temperature and related state parameters into consideration. Preferably, n=20 is taken from the constructed post-generator bearing temperature trend prediction model, i.e. the change of the relevant state parameter in 10 continuous minutes is considered; t=10 minutes, i.e. the post-generator bearing temperature is predicted after 10 minutes into the future.
Table 3 weight table of each status parameter
S134, carrying out smoothing treatment on the state parameter data related to the temperature of the rear bearing of the generator to obtain the state parameter related to the temperature of the rear bearing of the generator after the smoothing treatment. In this embodiment, a moving average method is selected to perform smoothing processing on the parameter data, where the smoothing processing method is as follows: considering the data point at the current moment and the previous h-1 data points in the time sequence, and taking the state parameter P related to the temperature fault of the bearing after the generator at the moment t into consideration i(t) The method comprises the following steps:
where the meaning of h is to consider h data points in the time series. Preferably, h=20, i.e. consider the operating state of the generator within 10 minutes.
S135, taking the post-generator bearing temperature at N continuous moments and the post-generator bearing temperature related state parameters after smoothing treatment at N continuous moments in the time sequence as the input of an SVR model, taking the residual time from a fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model.
In one embodiment of the present invention, there is also provided a post-generator bearing temperature fault prediction system 200, the post-generator bearing temperature fault prediction system 200 comprising: a data acquisition module 210, configured to acquire data of a fan operation state parameter, where the data includes historical data and current data; the related state parameter acquisition module 220 is connected with the data acquisition module and is used for acquiring state parameters related to the temperature of the rear bearing of the generator according to the data; the model construction module 230 is connected with the relevant state parameter acquisition module 220 and is used for constructing a post-generator bearing temperature prediction model based on an SVR model and/or a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model; the prediction module 240 is connected to the model building module 230 and the data acquisition module 210, respectively, and is configured to predict the bearing temperature after the generator and the remaining time from the occurrence of the bearing temperature alarm fault after the generator. Referring to fig. 5, a schematic structural diagram of an implementation of a method for predicting a temperature failure of a rear bearing of a generator according to an embodiment of the invention is shown.
In this embodiment, the data obtaining module 210 is configured to obtain data of the fan operation status parameter, where the data includes historical data and current data. Specifically, the data acquisition module 210 can directly invoke all parameters in the SCADA system regarding the operational status of the blower. The data acquisition module 210 not only selects the current value data of the state parameters, but also takes into account the historical data of the state parameters to enhance the accuracy of the prediction results, taking into account the timing of the post-generator bearing temperature and the related state parameters.
In this embodiment, the related state parameter obtaining module 220 is connected to the data obtaining module 210, and can select a state parameter related to the temperature of the rear bearing of the generator according to the fan operation state parameter obtained from the data obtaining module 210.
In this embodiment, the model building module 230 can build a post-generator bearing temperature prediction model based on the SVR model and/or a post-generator bearing temperature alarm fault remaining time prediction model based on the SVR model by using the state parameters obtained by the related parameter obtaining module 220. Specifically, the model building module 230 divides the data set of the state parameters obtained by the related parameter obtaining module 220 into a test set and a training set according to a certain proportion, preferably, the proportion is 2:8, the test set data is used for testing prediction precision, and the training set data is used for training a prediction model; and then training the prediction model, training the SVR model by using the data set, and selecting the optimal system parameters to improve the prediction accuracy.
In this embodiment, the prediction module 240 is connected to the model building module 230 and the data obtaining module 210, respectively, and when the system monitors that the running state of the generator is degraded and sends an alarm, the prediction module 240 predicts the temperature of the rear bearing of the generator and/or the residual time from the occurrence of the failure of the temperature alarm of the rear bearing of the generator according to the temperature related state parameter data of the rear bearing of the online SCADA system, by using the above prediction model.
In one embodiment of the present invention, the model building module 230 includes: a parameter weight obtaining unit 231, configured to obtain a parameter weight of the state parameter related to the post-generator bearing temperature in the prediction model; a super parameter obtaining unit 232, configured to obtain an optimal super parameter; the smoothing processing unit 233 is configured to obtain a state parameter related to a temperature of the rear bearing of the generator after the smoothing processing; the temperature prediction unit 234 is respectively connected with the parameter weight acquisition unit and the super parameter acquisition unit, and is used for constructing the post-bearing temperature prediction model of the generator based on the SVR model; the residual fault time prediction unit 235 is respectively connected to the parameter weight acquisition unit, the super parameter acquisition unit and the smoothing processing unit, and is used for constructing the residual fault time prediction model of the bearing temperature alarm after the generator based on the SVR model. Referring to fig. 6, a schematic diagram of an implementation structure of the model building module according to the present embodiment is shown.
In an embodiment of the present invention, there is also provided a computer-readable storage medium having stored thereon a computer program, characterized in that the program when called by a processor implements the post-generator bearing temperature failure prediction method as described above.
In an embodiment of the present invention, there is further provided a controller, including: a memory for storing a computer program; and the processor is in communication connection with the memory and is used for realizing the method for predicting the temperature failure of the rear bearing of the generator when the computer program is called.
The protection scope of the method for predicting the temperature failure of the rear bearing of the generator is not limited to the execution sequence of the steps listed in the embodiment, and all the schemes of step increase and decrease and step replacement in the prior art according to the principles of the invention are included in the protection scope of the invention.
The invention also provides a system for predicting the temperature failure of the rear bearing of the generator, which can realize the method for predicting the temperature failure of the rear bearing of the generator, but the device for realizing the method for predicting the temperature failure of the rear bearing of the generator comprises, but is not limited to, the structure of the system for predicting the temperature failure of the rear bearing of the generator, which is listed in the embodiment, and all the structural deformation and replacement of the prior art according to the principles of the invention are included in the protection scope of the invention.
The method, the system and the controller for predicting the temperature faults of the rear bearing of the motor have the following beneficial effects: firstly, in the design process, the invention calculates the correlation of all state parameters in an SCADA (Supervisory Control And Data Acquisition, data acquisition and monitoring control) system, and selects the state parameters with stronger correlation with the temperature of a bearing behind a generator from the calculated state parameters, thereby ensuring the comprehensiveness of the input parameters of the prediction system; secondly, according to the invention, the state parameters related to the temperature of the bearing behind the generator are selected through a correlation analysis method, the influence of each state parameter on the temperature of the bearing behind the generator is considered, the weight of each state parameter is designed according to the correlation analysis result, the parameter data is smoothed according to a moving smoothing method, the optimal super-parameters are searched in a mode of combining grid search and cross verification, the time sequence of the temperature of the bearing behind the generator and the related state parameters is considered, the temperature of the bearing behind the generator and the related state parameters at N continuous moments in the time sequence are taken as the input of a model, and a SVR model-based prediction model of the temperature of the bearing behind the generator and/or a prediction model of the residual time of alarm fault of the temperature of the bearing behind the generator are established, so that the prediction precision is higher; thirdly, the invention not only considers the prediction of the temperature of the rear bearing of the generator, but also predicts the residual time of the alarm fault of the temperature of the rear bearing of the generator, thereby providing guiding advice and assistance for on-site engineers to take effective and reasonable protective measures. In summary, the present invention effectively overcomes the disadvantages of the prior art and has high industrial utility value.
The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, it is intended that all equivalent modifications and variations of the invention be covered by the claims, which are within the ordinary skill of the art, be within the spirit and scope of the present disclosure.
Claims (8)
1. The method for predicting the temperature failure of the rear bearing of the generator is characterized by comprising the following steps of:
acquiring data of fan operation state parameters, wherein the data comprises historical data and current data;
acquiring state parameters related to the temperature of a rear bearing of the generator according to the data;
constructing a post-generator bearing temperature prediction model based on an SVR model and a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model by utilizing the state parameters;
predicting the temperature of the rear bearing of the generator and the residual time of the alarm fault of the temperature of the rear bearing of the generator;
the construction of the SVR model-based bearing temperature alarm fault remaining time prediction model after the generator by utilizing the state parameters comprises the following steps:
acquiring parameter weights of the state parameters related to the temperature of the rear bearing of the generator in the prediction model;
obtaining optimal super parameters by adopting a mode of combining grid search and cross verification;
carrying out smoothing treatment on the state parameter data related to the temperature of the rear bearing of the generator to obtain the state parameter related to the temperature of the rear bearing of the generator after the smoothing treatment, wherein the smoothing treatment method comprises the following steps: considering the data point at the current moment and the previous h-1 data points in the time sequence, and taking the state parameter P related to the temperature fault of the bearing after the generator at the moment t into consideration i(t) Is arranged asWherein the meaning of h is h data points in the considered time sequence;
and taking the state parameters related to the temperature of the rear bearing of the generator after the smoothing treatment as the input of an SVR model, taking the residual time of the distance fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct a residual time prediction model of the temperature alarm fault of the rear bearing of the generator based on the SVR model.
2. The post-generator bearing temperature fault prediction method according to claim 1, wherein the implementation manner of acquiring the state parameter related to the post-generator bearing temperature according to the data comprises:
analyzing the data by using a correlation analysis method and obtaining a correlation analysis result;
and selecting a state parameter with the correlation larger than a correlation threshold value in the correlation analysis result as a state parameter related to the temperature of the rear bearing of the generator.
3. The post-generator bearing temperature fault prediction method according to claim 2, wherein the correlation coefficients employed in the correlation analysis method include Pearson correlation coefficients and maximum information coefficients.
4. The post-generator bearing temperature fault prediction method according to claim 1, wherein the implementation manner of constructing a post-generator bearing temperature prediction model based on an SVR model and a post-generator bearing temperature alarm fault remaining time prediction model based on the SVR model by using the state parameters comprises:
acquiring parameter weights of the state parameters related to the temperature of the rear bearing of the generator in the prediction model;
obtaining optimal super parameters by adopting a mode of combining grid search and cross verification;
and taking the temperature of the generator rear bearing at the continuous N moments and the state parameters related to the temperature of the generator rear bearing at the continuous N moments in time sequence as the input of the SVR model, wherein N is a positive integer, the temperature of the generator rear bearing at the future T moment is taken as the output, and combining the parameter weight and the optimal super parameter to construct the SVR model-based generator rear bearing temperature prediction model.
5. The post-generator bearing temperature fault prediction method according to claim 4, wherein the obtaining the parameter weights of the state parameters related to the post-generator bearing temperature in the prediction model is implemented by:
for the ith state parameter, the parameter weight coefficientThe method comprises the following steps:
w i ＝r i ×MIC i
wherein w is i For a calculated intermediate value, r i Representing the samePearson correlation coefficient, MIC (MIC) of state parameter and rear bearing temperature of generator i The maximum information coefficient value representing the state parameter and the temperature of the bearing behind the generator, M represents the number of relevant state parameters.
6. A post-generator bearing temperature fault prediction system, the post-generator bearing temperature fault prediction system comprising:
the data acquisition module is used for acquiring data of the running state parameters of the fan, wherein the data comprise historical data and current data;
the related state parameter acquisition module is connected with the data acquisition module and is used for acquiring state parameters related to the temperature of the rear bearing of the generator according to the data;
the model construction module is connected with the related state parameter acquisition module and is used for constructing a post-generator bearing temperature prediction model based on an SVR model and a post-generator bearing temperature alarm fault residual time prediction model based on the SVR model;
the prediction module is respectively connected with the model construction module and the data acquisition module and is used for predicting the temperature of the bearing behind the generator and the residual time of the bearing temperature alarm fault after the generator;
the model building module comprises a parameter weight acquisition unit, a super parameter acquisition unit, a smoothing unit, a temperature prediction unit and a fault remaining time prediction unit, wherein:
the parameter weight acquisition unit is used for acquiring the parameter weight of the state parameter related to the temperature of the rear bearing of the generator in the prediction model;
the super-parameter acquisition unit is used for acquiring optimal super-parameters by adopting a mode of combining grid search and cross verification;
the smoothing unit is used for carrying out smoothing on the state parameter data related to the temperature of the rear bearing of the generator to obtain the state parameter related to the temperature of the rear bearing of the generator after the smoothing, and the smoothing method comprises the following steps: consider the data point at the current time and the number of previous h-1 in the time sequenceThe point is that the state parameter P related to the temperature fault of the bearing after the generator at the moment t i(t) Is arranged asWherein the meaning of h is h data points in the considered time sequence;
the temperature prediction unit is respectively connected with the parameter weight acquisition unit and the super parameter acquisition unit and is used for constructing the SVR model-based generator rear bearing temperature prediction model;
the fault residual time prediction unit is respectively connected with the parameter weight acquisition unit, the super parameter acquisition unit and the smoothing processing unit and is used for taking the post-generator bearing temperature related state parameter after smoothing processing as the input of an SVR model, taking the residual time of a distance fault as the output of the model, and combining the parameter weight and the optimal super parameter to construct the post-generator bearing temperature alarm fault residual time prediction model based on the SVR model.
7. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when called by a processor, implements the post-generator bearing temperature fault prediction method according to any one of claims 1 to 5.
8. A controller, comprising:
a memory for storing a computer program;
a processor, in communication with the memory, for implementing the post-generator bearing temperature fault prediction method of any one of claims 1 to 5 when the computer program is invoked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910402698.2A CN110135064B (en) | 2019-05-15 | 2019-05-15 | Method, system and controller for predicting temperature faults of rear bearing of generator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910402698.2A CN110135064B (en) | 2019-05-15 | 2019-05-15 | Method, system and controller for predicting temperature faults of rear bearing of generator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110135064A CN110135064A (en) | 2019-08-16 |
| CN110135064B true CN110135064B (en) | 2023-07-18 |
Family
ID=67574206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910402698.2A Active CN110135064B (en) | 2019-05-15 | 2019-05-15 | Method, system and controller for predicting temperature faults of rear bearing of generator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110135064B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110685868A (en) * | 2019-10-25 | 2020-01-14 | 长沙理工大学 | Wind turbine generator fault detection method and device based on improved gradient elevator |
| CN110805534B (en) * | 2019-11-18 | 2021-02-12 | 长沙理工大学 | Fault detection method, device and equipment of wind driven generator |
| CN113052716B (en) * | 2019-12-27 | 2023-03-03 | 新疆金风科技股份有限公司 | Abnormity early warning method and device for main bearing of wind generating set |
| CN112001632B (en) * | 2020-08-25 | 2022-07-19 | 中国船舶重工集团海装风电股份有限公司 | Wind turbine generator gearbox bearing performance degradation state evaluation method |
| CN112798949A (en) * | 2020-10-22 | 2021-05-14 | 国家电网有限公司 | Pumped storage unit generator temperature early warning method and system |
| CN112682273B (en) * | 2020-12-31 | 2022-06-28 | 长沙理工大学 | Wind turbine generator fault detection method based on cost-sensitive lightweight gradient elevator |
| CN113114748A (en) * | 2021-04-02 | 2021-07-13 | 天津一重电气自动化有限公司 | Intelligent prediction system based on data analysis |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951851A (en) * | 2015-07-08 | 2015-09-30 | 华侨大学 | Wind turbine state prediction model establishing method based on grey relation-regression SVM (support vector machine) |
| CN106769051A (en) * | 2017-03-10 | 2017-05-31 | 哈尔滨理工大学 | A kind of rolling bearing remaining life Forecasting Methodology based on MCEA KPCA and combination S VR |
| CN107977508A (en) * | 2017-11-29 | 2018-05-01 | 北京优利康达科技股份有限公司 | A kind of dynamo bearing failure prediction method |
| CN108680358A (en) * | 2018-03-23 | 2018-10-19 | 河海大学 | A kind of Wind turbines failure prediction method based on bearing temperature model |
| CN109389143A (en) * | 2018-06-19 | 2019-02-26 | 北京九章云极科技有限公司 | A kind of Data Analysis Services system and method for automatic modeling |
| CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
-
2019
- 2019-05-15 CN CN201910402698.2A patent/CN110135064B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951851A (en) * | 2015-07-08 | 2015-09-30 | 华侨大学 | Wind turbine state prediction model establishing method based on grey relation-regression SVM (support vector machine) |
| CN106769051A (en) * | 2017-03-10 | 2017-05-31 | 哈尔滨理工大学 | A kind of rolling bearing remaining life Forecasting Methodology based on MCEA KPCA and combination S VR |
| CN107977508A (en) * | 2017-11-29 | 2018-05-01 | 北京优利康达科技股份有限公司 | A kind of dynamo bearing failure prediction method |
| CN108680358A (en) * | 2018-03-23 | 2018-10-19 | 河海大学 | A kind of Wind turbines failure prediction method based on bearing temperature model |
| CN109389143A (en) * | 2018-06-19 | 2019-02-26 | 北京九章云极科技有限公司 | A kind of Data Analysis Services system and method for automatic modeling |
| CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
Non-Patent Citations (1)
| Title |
|---|
| 基于支持向量机和统计过程控制的风电机组故障预测算法的研究;胡庆春;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20120415;C042-259 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110135064A (en) | 2019-08-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110135064B (en) | Method, system and controller for predicting temperature faults of rear bearing of generator | |
| Wang et al. | Wind turbine fault detection and identification through PCA-based optimal variable selection | |
| Wang et al. | Fault prognostics using dynamic wavelet neural networks | |
| Lv et al. | An early fault detection method for induced draft fans based on MSET with informative memory matrix selection | |
| CN112287504A (en) | Offline/online integrated simulation system and method for power distribution network | |
| CN112069727B (en) | Intelligent transient stability evaluation system and method with high reliability for power system | |
| Niu et al. | Lebesgue sampling based deep belief network for lithium-ion battery diagnosis and prognosis | |
| CN115562225A (en) | Industrial robot operation and maintenance management method and device, computer equipment and storage medium | |
| CN115453356A (en) | Power equipment running state monitoring and analyzing method, system, terminal and medium | |
| CN114784882A (en) | Unit combination optimization processing method and device | |
| Khan et al. | Digital twin for advanced automation of future smart grid | |
| Baboli et al. | Two-stage condition-based maintenance model of wind turbine: From diagnosis to prognosis | |
| Zhao et al. | Gas turbine exhaust system health management based on recurrent neural networks | |
| CN108074045A (en) | The vulnerability analysis of Wind turbines complete machine and fault sequencing method and electric terminal | |
| Huang et al. | Elman neural network considering dynamic time delay estimation for short-term forecasting of offshore wind power | |
| Huang et al. | Digital twin modeling and operating state assessment of substation equipment | |
| Bourguet et al. | Artificial neural networks in electric power industry | |
| CN109739210B (en) | Method and device for evaluating health state of equipment component | |
| Li et al. | Predicting remaining useful life of industrial equipment based on multivariable monitoring data analysis | |
| CN114330865A (en) | Power grid reserve capacity prediction method and system, computer equipment and storage medium | |
| Yang et al. | Multistate prediction for in-service gas turbine via adaptive diffusion graph network | |
| Ou et al. | Fault Prediction Model of Wind Power Pitch System Based on BP Neural Network | |
| CN117728585B (en) | Distributed photovoltaic power station state prediction method and system | |
| CN117034157B (en) | Hydropower equipment fault identification method and system combining multimodal operation data | |
| Fang et al. | Failure Prediction of Wind Turbines Using Improved Gray Relation Analysis Based Support Vector Machine Method |
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 |