CN118013468B - Method, device, equipment and medium for monitoring health degree of wind turbine generator component - Google Patents
Method, device, equipment and medium for monitoring health degree of wind turbine generator component Download PDFInfo
- Publication number
- CN118013468B CN118013468B CN202410401645.XA CN202410401645A CN118013468B CN 118013468 B CN118013468 B CN 118013468B CN 202410401645 A CN202410401645 A CN 202410401645A CN 118013468 B CN118013468 B CN 118013468B
- Authority
- CN
- China
- Prior art keywords
- wind turbine
- regression model
- health
- health degree
- preset
- 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
- 230000036541 health Effects 0.000 title claims abstract description 198
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012544 monitoring process Methods 0.000 title claims abstract description 29
- 238000004364 calculation method Methods 0.000 claims abstract description 55
- 238000012549 training Methods 0.000 claims abstract description 46
- 230000006870 function Effects 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000012806 monitoring device Methods 0.000 claims 1
- 238000012423 maintenance Methods 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000010248 power generation Methods 0.000 abstract description 2
- 239000013598 vector Substances 0.000 description 10
- 238000011156 evaluation Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 238000009826 distribution Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 3
- 238000006731 degradation reaction Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
Landscapes
- Wind Motors (AREA)
Abstract
The application discloses a method, a device, equipment and a medium for monitoring the health degree of a wind turbine generator component, which relate to the technical field of equipment detection and comprise the following steps: constructing a sample data set based on wind turbine generator data, and training a temperature threshold calculation regression model by using the sample data set to obtain a trained regression model; determining a plurality of target regression models corresponding to a plurality of preset quantiles based on the trained regression models, so as to determine a plurality of component health reference values through the plurality of target regression models; and constructing a health degree scoring formula based on the health degree reference values of the components, and determining health degree scores of the components according to the health degree scoring formula and the real-time temperature of the wind turbine generator components so as to determine whether to automatically alarm according to the scores. Therefore, standard alignment analysis of the same component operation condition of the whole-farm unit can be carried out by analyzing parameters such as component temperature in the wind power station, component health index is calculated, fault trend early warning information is extracted, and intelligent operation and maintenance management level of wind power generation is improved.
Description
Technical Field
The invention relates to the technical field of equipment detection, in particular to a method, a device, equipment and a medium for monitoring the health of a wind turbine generator component.
Background
The wind generating set is generally built in a geographical position with good wind resources, is mainly distributed in hills and plains, has the characteristic of inconvenient traffic, and meanwhile, the cabin of the wind generating set is relatively narrow, so that the problem of difficult operation of maintenance personnel is caused. The large components such as a gear box, a generator, a frequency converter and the like in the wind turbine generator are key to fault diagnosis, and the power generation efficiency is reduced due to the fact that the temperature of the related components is too high.
The existing fault early warning method based on the component temperature mainly utilizes historical data and other measuring point data to fit the temperature of a key component, and judges whether an alarm is given or not by comparing the residual errors of a fitting value and an actual value with a threshold value, and the main defect is that only single unit component data is adopted, so that the fault condition is difficult to accurately judge under the condition that the degradation trend is not obvious for the existing potential fault gearbox. Meanwhile, part of exogenous variables cannot be acquired through measuring points at a station side, wherein part of required data is missing in the measuring points of SCADA (Supervisory Control And Data Acquisition, data acquisition and monitoring control system) due to non-uniform brand and model of a unit, data access line, coding problems and the like, and the cost of modes such as adding sensors and the like is high, and at the moment, the problem of low model accuracy is caused by information missing caused by the fact that a model is not fully input by using single unit data for diagnosis.
Disclosure of Invention
In view of the above, the invention aims to provide a method, a device, equipment and a medium for monitoring the health of wind turbine generator components, which can utilize the similarity of external environments of the same station units to perform calibration on the station units, transversely compare the output states of the units under the condition that constraint conditions such as model numbers and ambient temperature of the units are the same, determine the units with poor output by a calibration mode among the units, extract health indexes, and perform early warning when the indexes are lower than a threshold value, thereby realizing critical component fault early warning. The specific scheme is as follows:
In a first aspect, the application discloses a method for monitoring the health of a wind turbine component, which comprises the following steps:
Constructing a sample data set based on wind turbine data of all the wind turbine sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model;
determining a plurality of target regression models corresponding to a plurality of preset scores by utilizing the trained regression models, so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models;
Constructing a health degree scoring formula through the health degree reference values of the components, and determining health degree scores of the wind turbine components according to the health degree scoring formula and the real-time temperatures of the wind turbine components;
and determining whether to automatically alarm according to the health degree score of the wind turbine generator.
Optionally, the constructing a sample data set based on wind turbine data of all the wind turbine groups of the wind farm station in the preset time range includes:
Traversing a plurality of groups of wind turbine data in a first preset time range of all the wind turbine groups of the wind power station through a uniform resource positioning system;
Time synchronization is carried out on the plurality of groups of wind turbine generator data based on the time stamp, and missing values and abnormal values in the plurality of groups of wind turbine generator data are removed, so that a plurality of groups of processed wind turbine generator data are obtained;
And removing the unit numbers of the plurality of groups of processed wind turbine data, summarizing the plurality of groups of processed wind turbine data to obtain summarized wind turbine data, and constructing a sample data set according to the summarized wind turbine data.
Optionally, training the constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model, including:
setting the quantile of the constructed temperature threshold calculation regression model to be 0.5, and training the temperature threshold calculation regression model by utilizing the sample data set to obtain a regression model to be determined;
Determining the fitting degree of the regression model to be determined and the temperature parameter variation trend according to the decision coefficient, so as to determine whether the model precision of the regression model to be determined reaches a preset model precision threshold value or not based on the fitting degree;
And if the model precision of the regression model to be determined does not reach the preset model precision threshold, carrying out parameter adjustment on the regression model to be determined and retraining until the model precision of the regression model to be determined reaches the preset model precision threshold so as to determine the regression model to be determined as a trained regression model.
Optionally, the training the temperature threshold calculation regression model using the sample dataset includes:
and training the temperature threshold calculation regression model based on the sample data set, and carrying out minimum solving on an absolute value residual error loss function in an asymmetric form so as to fit the temperature parameter change trend.
Optionally, the determining, by using the trained regression model, a plurality of target regression models corresponding to a plurality of preset scores, so as to determine, by using the plurality of target regression models, a plurality of component health reference values corresponding to the plurality of preset scores, includes:
Setting the quantiles of the trained regression model to be 0.1, 0.4, 0.6 and 0.9 respectively, and training the temperature threshold calculation regression models corresponding to the 0.1, 0.4, 0.6 and 0.9 quantiles by utilizing the sample data set respectively to obtain a first target regression model corresponding to the 0.1 quantile, a second target regression model corresponding to the 0.4 quantile, a third target regression model corresponding to the 0.6 quantile and a fourth target regression model corresponding to the 0.9 quantile;
And constructing a single-unit data set based on single-unit data in a preset second time range, and respectively inputting the single-unit data set into the first target regression model, the second target regression model, the third target regression model and the fourth target regression model to obtain corresponding first reference value, second reference value, third reference value and fourth reference value.
Optionally, the constructing a health degree scoring formula according to the health degree reference values of the components, and determining a health degree score of the wind turbine component according to the health degree scoring formula and the real-time temperature of the wind turbine component, includes:
Constructing a health degree scoring formula based on the plurality of component health degree reference values and a preset inverse proportion function or a preset logic function, wherein the health degree scoring formula is as follows:
；
Wherein score is a health score, Y predict is a real-time temperature, Y 90% is a fourth reference value, Y 60% is a third reference value, Y 40% is a second reference value, and Y 10% is a first reference value;
And determining the health degree score of the wind turbine component according to the health degree score formula and the real-time temperature of the wind turbine component.
Optionally, before determining whether to automatically alarm according to the health degree score of the wind turbine generator, the method further includes:
Constructing a health gradient based on the health score to set different health states for different health ranges;
correspondingly, the determining whether to automatically alarm according to the health degree score of the wind turbine generator set comprises the following steps:
and if the health state corresponding to the health degree range of the wind turbine generator set where the health degree score is located is a non-health state, performing automatic alarm operation.
In a second aspect, the application discloses a device for monitoring the health of a wind turbine component, which comprises:
The model training module is used for constructing a sample data set based on wind turbine generator set data of all wind turbine generator sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and carrying out parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model;
The reference value determining module is used for determining a plurality of target regression models corresponding to a plurality of preset scores by utilizing the trained regression models so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models;
the health degree score calculation module is used for constructing a health degree score formula through the health degree reference values of the components and determining health degree scores of the wind turbine components according to the health degree score formula and the real-time temperatures of the wind turbine components;
and the automatic alarm module is used for determining whether to automatically alarm according to the health degree score of the wind turbine generator.
In a third aspect, the present application discloses an electronic device, comprising:
A memory for storing a computer program;
And the processor is used for executing the computer program to realize the method for monitoring the health of the wind turbine generator component.
In a fourth aspect, the present application discloses a computer readable storage medium for storing a computer program which, when executed by a processor, implements a method for monitoring the health of a wind turbine component as described above.
Firstly, constructing a sample data set based on wind turbine generator data of all wind turbine generator sets of a wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model; determining a plurality of target regression models corresponding to a plurality of preset scores by using the trained regression models, determining a plurality of component health reference values corresponding to the plurality of preset scores by using the plurality of target regression models, constructing a health score formula by using the plurality of component health reference values, and determining health scores of wind turbine components according to the health score formula and real-time temperatures of the wind turbine components; and finally, determining whether to automatically alarm according to the health degree score of the wind turbine generator. Therefore, according to the method, a data set is required to be built based on collected wind turbine generator system data of all wind turbine generator systems of the wind power station within a preset time range, a regression model is trained by using the preset data set, and a plurality of component health reference values corresponding to a plurality of preset scores are determined based on the trained regression model; and constructing a health degree scoring formula according to the health degree reference values of the components, determining health degree scores of the wind turbine generator components based on real-time temperatures of the components by using the scoring formula, and determining whether to automatically alarm according to the health degree scores. Therefore, the similarity of the external environments of the same station units can be utilized to calibrate the station units, the output states of the station units are transversely compared under the condition that constraint conditions such as the model number and the ambient temperature of the station units are the same, the unit with poor output is determined in a calibration mode among the station units, the health index is extracted, early warning is carried out when the index is lower than a threshold value, and the fault early warning of key components is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for monitoring the health of a wind turbine component;
FIG. 2 is a flowchart of a method for monitoring health of a wind turbine component according to the present disclosure;
FIG. 3 is a flowchart of another specific method for monitoring the health of wind turbine components disclosed in the present application;
fig. 4 and fig. 5 are schematic diagrams of a health degree scoring function according to the present disclosure;
FIG. 6 is a schematic structural diagram of a device for monitoring the health of wind turbine components according to the present application;
fig. 7 is a block diagram of an electronic device according to the present disclosure.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The existing fault early warning method based on the component temperature has the main defects that only single unit component data is adopted, and the fault condition is difficult to accurately judge on the condition that the degradation trend is not obvious for the key components with potential faults.
In order to overcome the technical problems, the application discloses a method, a device, equipment and a medium for monitoring the health of wind turbine generator components, which can be used for calibrating a station unit by utilizing the similarity of external environments of the same station unit, and can be used for transversely comparing the output states of the unit under the condition that the constraint conditions such as the model number, the ambient temperature and the like of the unit are the same, determining the unit with poor output by a calibration mode among the units, extracting health indexes, and carrying out early warning when the indexes are lower than a threshold value so as to realize critical component fault early warning.
Referring to fig. 1, the embodiment of the invention discloses a method for monitoring the health of a wind turbine generator component, which comprises the following steps:
And S11, constructing a sample data set based on wind turbine generator data of all the wind turbine generator sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model.
In this embodiment, a uniform resource positioning system (Uniform Resource Locator, URL) is used to traverse several groups of wind turbine data within a first preset time range of all the wind turbine data of the wind farm, and it is to be noted that, in this embodiment, the first preset time range is set to 21 days, that is, the data such as wind speed, active power, ambient temperature, critical component temperature, etc. of 21 days in the past of each wind turbine need to be obtained by traversing the whole wind turbine data of the URL, and then the several groups of wind turbine data are paired based on the time stamp, and missing values and abnormal values in the several groups of wind turbine data are removed, so as to obtain several groups of processed wind turbine data.
It is further to be noted that after all the data are preprocessed, different preprocessed set data need to be summarized, specifically, set numbers of a plurality of sets of processed wind turbine data need to be removed, summarized wind turbine data are obtained, and a sample data set is constructed according to the summarized wind turbine data.
Therefore, the health degree calibration can be carried out through the whole-field unit data, the limitation that the traditional method utilizes single unit data to evaluate is eliminated, and the influence of sporadic factors on the health degree scoring is obviously reduced by utilizing the characteristics of the same unit operation characteristic and similar external environment of the whole field.
In this embodiment, the constructed temperature threshold calculation regression model is further required to be trained by using the sample data set, and parameters of the temperature threshold calculation regression model are adjusted to obtain a trained regression model, specifically, the quantile of the constructed temperature threshold calculation regression model is required to be set to 0.5 for training, and the fitting degree of the model and the temperature parameter variation trend is determined according to the decision coefficient, so that the accuracy of the trained regression model reaches the standard.
And S12, determining a plurality of target regression models corresponding to a plurality of preset scores by utilizing the trained regression models so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models.
In this embodiment, after determining the regression model after training, the data of the whole field unit in the past 21 days is required to be used as a training set, and the regression models of 0.1 quantile, 0.4 quantile, 0.6 quantile and 0.9 quantile are selected as the benchmark reference values of the component health evaluation threshold. And then, single-set data of 7 days is selected as a verification set, verification set data is applied to four boundary threshold calculation models, the condition distribution situation obtained based on the quantile regression model is calculated for each sample point, and four corresponding reference values are respectively obtained according to the four boundary threshold calculation models.
And S13, constructing a health degree scoring formula through the health degree reference values of the components, and determining health degree scores of the wind turbine components according to the health degree scoring formula and the real-time temperatures of the wind turbine components.
In this embodiment, a health degree scoring formula needs to be built according to the obtained reference value, and then the real-time temperature of the wind turbine component is used as an input value to calculate the health degree score of the wind turbine component through the formula. It should be noted that, according to the actual situation, a corresponding health degree gradient may be constructed to set different health degree states for different health degree ranges. For example, a component score of 80 points or more is defined as a healthy state, a component score of 60 points to 80 points is defined as a sub-healthy state, and a component score of 60 points or less is defined as unhealthy. The temperature distribution conditions of the same components of the full-field unit are compared, the operation conditions of the components are evaluated in a comparison mode, the station level benchmarking management is facilitated, meanwhile, a visual display means is formed for the benchmarking tamping basis of the health degree of the regional and regional level equipment.
And S14, determining whether to automatically alarm according to the health degree score of the wind turbine generator.
In the embodiment, based on an automatic alarm function of the health degree of the fan component, a special training list and an overhaul suggestion can be automatically generated according to the marking result of the health degree, and an effective guidance suggestion is provided for overhaul work of a power production site. The health level of the whole field unit can be marked, the unit scoring can be carried out in operation and maintenance management, the maintenance effect of field personnel is improved, the operation management effect of enterprises is enhanced, and the digital transformation of the power system is supported.
In the embodiment, a data set is required to be built based on collected wind turbine generator set data of all wind turbine generator sets of a wind power station within a preset time range, then a regression model is trained by using the preset data set, and a plurality of component health reference values corresponding to a plurality of preset quantiles are determined based on the trained regression model; and constructing a health degree scoring formula according to the health degree reference values of the components, determining health degree scores of the wind turbine generator components based on real-time temperatures of the components by using the scoring formula, and determining whether to automatically alarm according to the health degree scores. Therefore, the similarity of the external environments of the same station units can be utilized to calibrate the station units, the output states of the station units are transversely compared under the condition that constraint conditions such as the model number and the ambient temperature of the station units are the same, the unit with poor output is determined in a calibration mode among the station units, the health index is extracted, early warning is carried out when the index is lower than a threshold value, and the fault early warning of key components is realized.
Based on the foregoing embodiment, it can be seen that the training of the regression model calculated by the temperature threshold constructed earlier is required in the present application, for this reason, the embodiment of the present application describes in detail how to train the model, as shown in fig. 2, and the embodiment of the present application discloses a method for monitoring the health of a wind turbine component, including:
And S21, setting the quantile of the constructed temperature threshold calculation regression model to be 0.5, and training the temperature threshold calculation regression model by using a sample data set to obtain the regression model to be determined.
In this embodiment, the quantile of the constructed temperature threshold calculation regression model is set to 0.5, and the sample dataset is used to train the temperature threshold calculation regression model to obtain the regression model to be determined. Setting the q-ary fraction y q (x) of the y-to-x conditional distribution as a linear function of x, the formula is:
；
where x i is the set of input variable vectors, y q (x) is the set of target variable vectors, and β q is the bias vector.
Estimation of the quantityThe absolute residual loss function (loss q) in an asymmetric form at q quantiles can be minimized, as follows:
；
In order to monitor the variation trend of the quantile model and ensure the fitting degree of the model, a q value of 0.5 is selected as the quantile model to monitor the variation form of the whole quantile model, and the q value of 0.5 is also called median regression, and the formula can be simplified as follows:
；
Wherein y i is the set of target variable vectors in the sample data, and beta q is the one required under the condition of minimum loss function 。
It should be noted that the median regression is called a minimum absolute difference estimator (Least Absolute Deviation Estimator). Compared with the traditional linear regression, the square of absolute value residual error is taken as an estimated quantity, the influence of abnormal points on quantile regression is remarkably reduced, error items are not required to be normally distributed, good robustness is achieved, meanwhile, the overall view of the condition distribution of the explained variable can be relatively comprehensively described, the condition expectation of the explained variable and the quantile of the explained variable influenced by the condition expectation can be analyzed, trend lines established by the quantile regression are provided, and visual analysis and system understanding of the model are facilitated.
And S22, determining the fitting degree of the regression model to be determined and the temperature parameter change trend according to the decision coefficient so as to determine whether the model precision of the regression model to be determined reaches a preset model precision threshold value or not based on the fitting degree.
In this embodiment, after the median model is trained, the fitting condition of the model is actively tracked, and the decision coefficient R 2 is adopted to perform evaluation and assessment so as to quantify the degree of deviation of the model from the data in the test set. The decision coefficients, also called deterministic coefficients, are commonly used to evaluate the goodness of a linear fit. In the step, the accuracy verification is carried out on the median regression model by adopting the decision coefficient, and the accuracy and the universality of the model are proved by selecting test set data. To calculate the decision coefficients, the square of the change of the observed value of the target variable is defined as the sum of squares of regression (SSR, regression sum of squares), the square of the total change of the actual value of the target variable is defined as the sum of squares of total deviation (SST, total sum of square), and the formula is:
；
；
wherein, X is input variable vector group, y is output variable vector group, b is bias vector, M 0 is idempotent and symmetrical, which is used for calculating variance of output variable in training data set, and the formula is as follows:
；
Wherein i is a row vector with length n and 1 for each element, and i' is a transposed vector of i, i.e., a row vector with length n and 1 for each element.
The decision coefficients need to be calculated based on the sum of squares of the regression and the sum of squares of the total deviation, and the numerical value is used for deducing the proportion of the change (SSR) of the observed value of the target variable to the total change (SST) of the actual value of the target variable for the interpretation variable. The formula is as follows:
；
According to the formula, the determination coefficient can determine the deviation degree of the model through the square of the error accounting for the total change proportion of the target variable, and the calculation model can explain the evaluation parameter of the actual change percentage. And comparing the fitting goodness threshold with the calculated decision coefficient, and ensuring the accuracy of model training by repeatedly debugging training parameters and a better data cleaning means.
And S23, if the model precision of the regression model to be determined does not reach the preset model precision threshold, carrying out parameter adjustment on the regression model to be determined, and retraining until the model precision of the regression model to be determined reaches the preset model precision threshold, so as to determine the regression model to be determined as a trained regression model.
In this embodiment, if the model accuracy of the regression model to be determined does not reach the preset model accuracy threshold, after the model parameters are adjusted, training the model is continued until the model accuracy reaches the preset model accuracy threshold, if the model accuracy of the regression model to be determined reaches the preset model accuracy threshold, the characterization model and the temperature parameters are already fitted, and the regression model to be determined can be determined as a trained regression model.
Therefore, compared with the traditional linear regression, the method provided by the application has the advantages that the quantile regression is adopted on model training, the quantile regression is obviously reduced by the influence of abnormal points, the error items are not required to be normally distributed, the method has good robustness, meanwhile, the overall view of the condition distribution of the explained variable can be relatively comprehensively described, the condition expectation of the explained variable and the quantile number of the explained variable influenced by the condition expectation can be analyzed, the trend line established by each quantile regression is provided, and the visual analysis and the system understanding of the model are facilitated. And the fitting goodness threshold value is adopted to be compared with the calculated decision coefficient, and the accuracy of model training is ensured by repeatedly debugging training parameters and a better data cleaning means.
Based on the foregoing embodiments, according to the method of the present application, it is necessary to construct a health degree scoring formula and perform health degree scoring according to the health degree scoring formula. Referring to fig. 3, the embodiment of the application discloses a method for monitoring the health of a wind turbine generator component, which comprises the following steps:
And S31, constructing a sample data set based on wind turbine generator data of all the wind turbine generator sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model.
And S32, respectively setting the quantiles of the trained regression model to be 0.1, 0.4, 0.6 and 0.9, and respectively training the temperature threshold calculation regression models corresponding to the 0.1, 0.4, 0.6 and 0.9 quantiles by utilizing the sample data set to obtain a first target regression model corresponding to the 0.1 quantile, a second target regression model corresponding to the 0.4 quantile, a third target regression model corresponding to the 0.6 quantile and a fourth target regression model corresponding to the 0.9 quantile.
In this embodiment, when the accuracy of the regression model after training reaches the standard, the data of the whole field unit in the past 21 days may be used as a training set, and the regression model with 0.1 quantile, 0.4 quantile, 0.6 quantile and 0.9 quantile is selected as the benchmark reference value of the component health evaluation threshold. The formula is as follows:
；
；
；
；
And step S33, a single-unit data set is constructed based on single-unit data in a preset second time range, and the single-unit data set is respectively input into the first target regression model, the second target regression model, the third target regression model and the fourth target regression model to obtain corresponding first reference value, second reference value, third reference value and fourth reference value.
In this embodiment, single-set data of the past 7 days is required to be selected as a verification set, the verification set data is applied to four boundary threshold calculation models, the condition distribution situation obtained by calculating based on the quantile regression model is obtained for each sample point, and corresponding reference values Y 10%、Y40%、Y60%、Y90% are respectively obtained according to four formulas.
And step S34, constructing a health degree scoring formula through the health degree reference values of the components, and determining health degree scores of the wind turbine components according to the health degree scoring formula and the real-time temperatures of the wind turbine components.
In this embodiment, the composite score is calculated based on the degree of deviation. Y 10%、Y40%、Y60%、Y90% for a single sample point. As a reference value, a piecewise function is used in combination with a modulated inverse proportion function or a logic function to provide a component health degree scoring reference, and under consideration of seasonal influence, a referenceable threshold setting includes a score of 100 points in the actual value range Y 40% to Y 60% and a score of 60 points in the range Y 10%-0.2（Y40%-Y10%) -12 or Y 90%+0.2（Y90%-Y60%) +8, and the score calculation formula can be configured as follows:
；
Wherein score is a health score, Y predict is a real-time temperature, Y 90% is a fourth reference value, Y 60% is a third reference value, Y 40% is a second reference value, and Y 10% is a first reference value;
And then determining the health degree score of the wind turbine component according to the health degree score formula and the real-time temperature of the wind turbine component.
It should be noted that the function may be visually displayed, and examples thereof are shown in fig. 4 and 5.
Step S35, constructing a health degree gradient based on the health degree score so as to set different health degree states for different health degree ranges.
And S36, if the health state corresponding to the health degree range of the wind turbine generator is a non-health state, performing automatic alarm operation.
Therefore, in this embodiment, the benchmark reference value of the health evaluation threshold value can be determined according to the models corresponding to different scores, and then the health evaluation formula is constructed according to the obtained benchmark reference value, so that the degradation trend can be found in early stage by comprehensively evaluating the operation condition of the component in a health evaluation mode, the low-efficiency operation condition of the unit is exposed before the fault occurs, and the planned replacement of the key component is achieved.
Referring to fig. 6, an embodiment of the invention discloses a device for monitoring health of wind turbine components, which comprises:
The model training module 11 is configured to construct a sample data set based on wind turbine generator data of all wind turbine generator sets of the wind farm station within a preset time range, train the constructed temperature threshold calculation regression model by using the sample data set, and perform parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model;
A reference value determining module 12, configured to determine a plurality of target regression models corresponding to a plurality of preset scores by using the trained regression models, so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models;
The health degree score calculation module 13 is configured to construct a health degree score formula according to the health degree reference values of the plurality of components, and determine a health degree score of the wind turbine component according to the health degree score formula and the real-time temperature of the wind turbine component;
And the automatic alarm module 14 is used for determining whether to automatically alarm according to the health degree score of the wind turbine generator.
In the embodiment, firstly, a sample data set is constructed based on wind turbine generator data of all wind turbine generator sets of a wind power station within a preset time range, the constructed temperature threshold calculation regression model is trained by using the sample data set, and parameter adjustment is performed on the temperature threshold calculation regression model to obtain a trained regression model; determining a plurality of target regression models corresponding to a plurality of preset scores by using the trained regression models, determining a plurality of component health reference values corresponding to the plurality of preset scores by using the plurality of target regression models, constructing a health score formula by using the plurality of component health reference values, and determining health scores of wind turbine components according to the health score formula and real-time temperatures of the wind turbine components; and finally, determining whether to automatically alarm according to the health degree score of the wind turbine generator. Therefore, according to the method, a data set is required to be built based on collected wind turbine generator system data of all wind turbine generator systems of the wind power station within a preset time range, a regression model is trained by using the preset data set, and a plurality of component health reference values corresponding to a plurality of preset scores are determined based on the trained regression model; and constructing a health degree scoring formula according to the health degree reference values of the components, determining health degree scores of the wind turbine generator components based on real-time temperatures of the components by using the scoring formula, and determining whether to automatically alarm according to the health degree scores. Therefore, the similarity of the external environments of the same station units can be utilized to calibrate the station units, the output states of the station units are transversely compared under the condition that constraint conditions such as the model number and the ambient temperature of the station units are the same, the unit with poor output is determined in a calibration mode among the station units, the health index is extracted, early warning is carried out when the index is lower than a threshold value, and the fault early warning of key components is realized.
In some embodiments, the model training module 11 may specifically include:
the data acquisition unit is used for traversing a plurality of groups of wind turbine generator data in a first preset time range of all the wind turbine generator sets of the wind power station through the uniform resource positioning system;
The data processing unit is used for timing the plurality of groups of wind turbine generator data based on the time stamp, and eliminating missing values and abnormal values in the plurality of groups of wind turbine generator data so as to obtain a plurality of groups of processed wind turbine generator data;
The data set construction unit is used for removing the unit numbers of the plurality of groups of processed wind turbine data, summarizing the plurality of groups of processed wind turbine data to obtain summarized wind turbine data, and constructing a sample data set according to the summarized wind turbine data.
In some embodiments, the model training module 11 may specifically include:
The model training submodule is used for setting the quantile of the constructed temperature threshold calculation regression model to be 0.5, and training the temperature threshold calculation regression model by utilizing the sample data set to obtain a regression model to be determined;
the accuracy determining submodule is used for determining the fitting degree of the regression model to be determined and the temperature parameter change trend according to the decision coefficient so as to determine whether the model accuracy of the regression model to be determined reaches a preset model accuracy threshold value or not based on the fitting degree;
And the model determination submodule is used for carrying out parameter adjustment on the regression model to be determined and retraining until the model precision of the regression model to be determined reaches the preset model precision threshold value so as to determine the regression model to be determined as a trained regression model if the model precision of the regression model to be determined does not reach the preset model precision threshold value.
In some embodiments, the model training sub-module may specifically include:
And the model training unit is used for training the temperature threshold calculation regression model based on the sample data set, and fitting the temperature parameter change trend by carrying out minimum solution on the absolute value residual error loss function in an asymmetric form.
In some embodiments, the reference value determining module 12 may specifically include:
The model determining unit is used for setting the quantiles of the trained regression model to be 0.1, 0.4, 0.6 and 0.9 respectively, and training the temperature threshold calculation regression models corresponding to the 0.1, 0.4, 0.6 and 0.9 quantiles by utilizing the sample data set respectively to obtain a first target regression model corresponding to the 0.1 quantile, a second target regression model corresponding to the 0.4 quantile, a third target regression model corresponding to the 0.6 quantile and a fourth target regression model corresponding to the 0.9 quantile;
The reference value determining unit is used for constructing a single-unit group data set based on single-unit group data in a preset second time range, and inputting the single-unit group data set into the first target regression model, the second target regression model, the third target regression model and the fourth target regression model respectively to obtain corresponding first reference value, second reference value, third reference value and fourth reference value.
In some embodiments, the automatic alarm module 14 may specifically include:
the gradient setting unit is used for constructing a health degree gradient based on the health degree score so as to set different health degree states for different health degree ranges;
and the alarm unit is used for carrying out automatic alarm operation if the health state corresponding to the health degree range of the health degree score of the wind turbine generator is a non-health state.
Further, the embodiment of the present application further discloses an electronic device, and fig. 7 is a block diagram of an electronic device 20 according to an exemplary embodiment, where the content of the figure is not to be considered as any limitation on the scope of use of the present application.
Fig. 7 is a schematic structural diagram of an electronic device 20 according to an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input output interface 25, and a communication bus 26. The memory 22 is configured to store a computer program, where the computer program is loaded and executed by the processor 21 to implement relevant steps in the method for monitoring the health of a wind turbine component disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in the present embodiment may be specifically an electronic computer.
In this embodiment, the power supply 23 is configured to provide an operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and an external device, and the communication protocol in which the communication interface is in compliance is any communication protocol applicable to the technical solution of the present application, which is not specifically limited herein; the input/output interface 25 is used for acquiring external input data or outputting external output data, and the specific interface type thereof may be selected according to the specific application requirement, which is not limited herein.
The memory 22 may be a carrier for storing resources, such as a read-only memory, a random access memory, a magnetic disk, or an optical disk, and the resources stored thereon may include an operating system 221, a computer program 222, and the like, and the storage may be temporary storage or permanent storage.
The operating system 221 is used for managing and controlling various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, netware, unix, linux, etc. The computer program 222 may further comprise a computer program capable of performing other specific tasks in addition to the computer program capable of performing the method of monitoring the health of a wind turbine component performed by the electronic device 20 as disclosed in any of the previous embodiments.
Further, the application also discloses a computer readable storage medium for storing a computer program; the method for monitoring the health of the wind turbine generator component is characterized in that the computer program is executed by a processor. For specific steps of the method, reference may be made to the corresponding contents disclosed in the foregoing embodiments, and no further description is given here.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, so that the same or similar parts between the embodiments are referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The foregoing has outlined rather broadly the more detailed description of the application in order that the detailed description of the application that follows may be better understood, and in order that the present principles and embodiments may be better understood; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present application, the present description should not be construed as limiting the present application in view of the above.
Claims (8)
1. A method for monitoring the health of a wind turbine component, comprising:
Constructing a sample data set based on wind turbine data of all the wind turbine sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model;
determining a plurality of target regression models corresponding to a plurality of preset scores by utilizing the trained regression models, so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models;
Constructing a health degree scoring formula through the health degree reference values of the components, and determining health degree scores of the wind turbine components according to the health degree scoring formula and the real-time temperatures of the wind turbine components;
determining whether to automatically alarm according to the health degree score of the wind turbine generator;
wherein the determining, by using the trained regression model, a plurality of target regression models corresponding to a plurality of preset scores to determine, by using the plurality of target regression models, a plurality of component health reference values corresponding to the plurality of preset scores includes:
Setting the quantiles of the trained regression model to be 0.1, 0.4, 0.6 and 0.9 respectively, and training the temperature threshold calculation regression models corresponding to the 0.1, 0.4, 0.6 and 0.9 quantiles by utilizing the sample data set respectively to obtain a first target regression model corresponding to the 0.1 quantile, a second target regression model corresponding to the 0.4 quantile, a third target regression model corresponding to the 0.6 quantile and a fourth target regression model corresponding to the 0.9 quantile;
A single-unit data set is built based on single-unit data in a preset second time range, and the single-unit data set is respectively input into the first target regression model, the second target regression model, the third target regression model and the fourth target regression model to obtain corresponding first reference value, second reference value, third reference value and fourth reference value;
the health degree scoring formula is constructed through the health degree reference values of the components, and the health degree score of the wind turbine component is determined according to the health degree scoring formula and the real-time temperature of the wind turbine component, and the method comprises the following steps:
Constructing a health degree scoring formula based on the plurality of component health degree reference values and a preset inverse proportion function or a preset logic function, wherein the health degree scoring formula is as follows:
；
Wherein score is a health score, Y predict is a real-time temperature, Y 90% is a fourth reference value, Y 60% is a third reference value, Y 40% is a second reference value, and Y 10% is a first reference value;
And determining the health degree score of the wind turbine component according to the health degree score formula and the real-time temperature of the wind turbine component.
2. The method for monitoring the health of wind turbine components according to claim 1, wherein the constructing a sample data set based on wind turbine data of all wind turbines of the wind farm in a preset time range comprises:
Traversing a plurality of groups of wind turbine data in a first preset time range of all the wind turbine groups of the wind power station through a uniform resource positioning system;
Time synchronization is carried out on the plurality of groups of wind turbine generator data based on the time stamp, and missing values and abnormal values in the plurality of groups of wind turbine generator data are removed, so that a plurality of groups of processed wind turbine generator data are obtained;
And removing the unit numbers of the plurality of groups of processed wind turbine data, summarizing the plurality of groups of processed wind turbine data to obtain summarized wind turbine data, and constructing a sample data set according to the summarized wind turbine data.
3. The method for monitoring the health of a wind turbine component according to claim 1, wherein the training the constructed temperature threshold calculation regression model by using the sample data set and performing parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model comprises:
setting the quantile of the constructed temperature threshold calculation regression model to be 0.5, and training the temperature threshold calculation regression model by utilizing the sample data set to obtain a regression model to be determined;
Determining the fitting degree of the regression model to be determined and the temperature parameter variation trend according to the decision coefficient, so as to determine whether the model precision of the regression model to be determined reaches a preset model precision threshold value or not based on the fitting degree;
And if the model precision of the regression model to be determined does not reach the preset model precision threshold, carrying out parameter adjustment on the regression model to be determined and retraining until the model precision of the regression model to be determined reaches the preset model precision threshold so as to determine the regression model to be determined as a trained regression model.
4. A method of monitoring health of a wind turbine component according to claim 3, wherein training the temperature threshold calculation regression model using the sample dataset comprises:
and training the temperature threshold calculation regression model based on the sample data set, and carrying out minimum solving on an absolute value residual error loss function in an asymmetric form so as to fit the temperature parameter change trend.
5. A method of monitoring the health of a wind turbine component according to any of claims 1 to 4, further comprising, prior to determining whether to automatically alert based on the health score of the wind turbine:
Constructing a health gradient based on the health score to set different health states for different health ranges;
correspondingly, the determining whether to automatically alarm according to the health degree score of the wind turbine generator set comprises the following steps:
and if the health state corresponding to the health degree range of the wind turbine generator set where the health degree score is located is a non-health state, performing automatic alarm operation.
6. Wind turbine generator system part health monitoring devices, characterized in that includes:
The model training module is used for constructing a sample data set based on wind turbine generator set data of all wind turbine generator sets of the wind power station within a preset time range, training a constructed temperature threshold calculation regression model by using the sample data set, and carrying out parameter adjustment on the temperature threshold calculation regression model to obtain a trained regression model;
The reference value determining module is used for determining a plurality of target regression models corresponding to a plurality of preset scores by utilizing the trained regression models so as to determine a plurality of component health reference values corresponding to the plurality of preset scores through the plurality of target regression models;
the health degree score calculation module is used for constructing a health degree score formula through the health degree reference values of the components and determining health degree scores of the wind turbine components according to the health degree score formula and the real-time temperatures of the wind turbine components;
the automatic alarm module is used for determining whether to automatically alarm according to the health degree score of the wind turbine generator;
Wherein the reference value determining module includes:
The model determining unit is used for setting the quantiles of the trained regression model to be 0.1, 0.4, 0.6 and 0.9 respectively, and training the temperature threshold calculation regression models corresponding to the 0.1, 0.4, 0.6 and 0.9 quantiles by utilizing the sample data set respectively to obtain a first target regression model corresponding to the 0.1 quantile, a second target regression model corresponding to the 0.4 quantile, a third target regression model corresponding to the 0.6 quantile and a fourth target regression model corresponding to the 0.9 quantile;
The reference value determining unit is used for constructing a single-unit group data set based on single-unit group data in a preset second time range, and inputting the single-unit group data set into the first target regression model, the second target regression model, the third target regression model and the fourth target regression model respectively to obtain corresponding first reference value, second reference value, third reference value and fourth reference value;
wherein, the health degree score calculation module is used for:
Constructing a health degree scoring formula based on the plurality of component health degree reference values and a preset inverse proportion function or a preset logic function, wherein the health degree scoring formula is as follows:
；
Wherein score is a health score, Y predict is a real-time temperature, Y 90% is a fourth reference value, Y 60% is a third reference value, Y 40% is a second reference value, and Y 10% is a first reference value;
And determining the health degree score of the wind turbine component according to the health degree score formula and the real-time temperature of the wind turbine component.
7. An electronic device, comprising:
A memory for storing a computer program;
A processor for executing the computer program to implement a wind turbine component health monitoring method as claimed in any one of claims 1 to 5.
8. A computer readable storage medium for storing a computer program which when executed by a processor implements a method of monitoring the health of a wind turbine component according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410401645.XA CN118013468B (en) | 2024-04-03 | 2024-04-03 | Method, device, equipment and medium for monitoring health degree of wind turbine generator component |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410401645.XA CN118013468B (en) | 2024-04-03 | 2024-04-03 | Method, device, equipment and medium for monitoring health degree of wind turbine generator component |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118013468A CN118013468A (en) | 2024-05-10 |
| CN118013468B true CN118013468B (en) | 2024-06-11 |
Family
ID=90948801
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410401645.XA Active CN118013468B (en) | 2024-04-03 | 2024-04-03 | Method, device, equipment and medium for monitoring health degree of wind turbine generator component |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118013468B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10288043B2 (en) * | 2014-11-18 | 2019-05-14 | Abb Schweiz Ag | Wind turbine condition monitoring method and system |
| CN111582392A (en) * | 2020-05-12 | 2020-08-25 | 河北工业大学 | Multi-working-condition health state online monitoring method for key components of wind turbine generator |
| CN112597696A (en) * | 2020-12-07 | 2021-04-02 | 国网江苏省电力有限公司 | Boiler four-pipe leakage early warning method based on extreme learning machine principle |
| CN112818601A (en) * | 2021-02-05 | 2021-05-18 | 河海大学 | Hydroelectric generating set health assessment method based on GA-BP neural network and error statistical analysis |
| CN114065510A (en) * | 2021-11-15 | 2022-02-18 | 国能国华(北京)燃气热电有限公司 | Fault early warning method and device for cooling tower fan and electronic equipment |
| CN115419558A (en) * | 2022-09-27 | 2022-12-02 | 许继电气股份有限公司 | Wind turbine generator running state evaluation method and device |
| CN116522065A (en) * | 2023-04-03 | 2023-08-01 | 未必然数据科技(北京)有限公司 | Coal mill health degree assessment method based on deep learning |
| CN116624343A (en) * | 2023-06-14 | 2023-08-22 | 华北电力大学 | Wind turbine generator tower abnormal vibration monitoring and health degree evaluation method and system |
| CN116664099A (en) * | 2023-04-25 | 2023-08-29 | 安徽海博智能科技有限责任公司 | Unmanned mine car health management method based on nonparametric regression |
| CN116831523A (en) * | 2023-06-12 | 2023-10-03 | 中南大学 | Alarm method, device, equipment and storage medium based on health monitoring |
| CN117390370A (en) * | 2023-09-27 | 2024-01-12 | 上海朋熙半导体有限公司 | Machine early warning method, device, equipment and readable medium based on health index |
| CN117454324A (en) * | 2023-11-17 | 2024-01-26 | 华能陈巴尔虎旗风力发电有限公司 | Anemoclinograph fault early warning method, system, electronic equipment and storage medium |
| CN117563184A (en) * | 2024-01-15 | 2024-02-20 | 东营昆宇电源科技有限公司 | Energy storage fire control system based on thing networking |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220309388A1 (en) * | 2021-03-29 | 2022-09-29 | Allstate Insurance Company | Systems and Methods for Classification and Time Series Calibration in Identity Health Analysis |
-
2024
- 2024-04-03 CN CN202410401645.XA patent/CN118013468B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10288043B2 (en) * | 2014-11-18 | 2019-05-14 | Abb Schweiz Ag | Wind turbine condition monitoring method and system |
| CN111582392A (en) * | 2020-05-12 | 2020-08-25 | 河北工业大学 | Multi-working-condition health state online monitoring method for key components of wind turbine generator |
| CN112597696A (en) * | 2020-12-07 | 2021-04-02 | 国网江苏省电力有限公司 | Boiler four-pipe leakage early warning method based on extreme learning machine principle |
| CN112818601A (en) * | 2021-02-05 | 2021-05-18 | 河海大学 | Hydroelectric generating set health assessment method based on GA-BP neural network and error statistical analysis |
| CN114065510A (en) * | 2021-11-15 | 2022-02-18 | 国能国华(北京)燃气热电有限公司 | Fault early warning method and device for cooling tower fan and electronic equipment |
| CN115419558A (en) * | 2022-09-27 | 2022-12-02 | 许继电气股份有限公司 | Wind turbine generator running state evaluation method and device |
| CN116522065A (en) * | 2023-04-03 | 2023-08-01 | 未必然数据科技(北京)有限公司 | Coal mill health degree assessment method based on deep learning |
| CN116664099A (en) * | 2023-04-25 | 2023-08-29 | 安徽海博智能科技有限责任公司 | Unmanned mine car health management method based on nonparametric regression |
| CN116831523A (en) * | 2023-06-12 | 2023-10-03 | 中南大学 | Alarm method, device, equipment and storage medium based on health monitoring |
| CN116624343A (en) * | 2023-06-14 | 2023-08-22 | 华北电力大学 | Wind turbine generator tower abnormal vibration monitoring and health degree evaluation method and system |
| CN117390370A (en) * | 2023-09-27 | 2024-01-12 | 上海朋熙半导体有限公司 | Machine early warning method, device, equipment and readable medium based on health index |
| CN117454324A (en) * | 2023-11-17 | 2024-01-26 | 华能陈巴尔虎旗风力发电有限公司 | Anemoclinograph fault early warning method, system, electronic equipment and storage medium |
| CN117563184A (en) * | 2024-01-15 | 2024-02-20 | 东营昆宇电源科技有限公司 | Energy storage fire control system based on thing networking |
Non-Patent Citations (1)
| Title |
|---|
| 基于深度学习理论的机械装备大数据健康监测方法;雷亚国;贾峰;周昕;林京;;机械工程学报;20151130;第51卷(第21期);第49-56页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118013468A (en) | 2024-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109947088B (en) | Equipment fault early warning system based on model full life cycle management | |
| JP6423546B2 (en) | Advanced data cleansing system and method | |
| JP6467186B2 (en) | System and method for detecting, correcting, and inspecting bad data in a data stream | |
| CN107622308B (en) | Power generation equipment parameter early warning method based on DBN (database-based network) | |
| US20120310597A1 (en) | Failure cause diagnosis system and method | |
| CN110008565A (en) | A kind of industrial process unusual service condition prediction technique based on operating parameter association analysis | |
| JP2009294969A (en) | Demand forecast method and demand forecast device | |
| KR101908865B1 (en) | Method for data quality analysis of observed temperature | |
| CN111458661A (en) | Power distribution network line variation relation diagnosis method, device and system | |
| CN117270514B (en) | Production process whole-flow fault detection method based on industrial Internet of things | |
| Kirchen et al. | Metrics for the evaluation of data quality of signal data in industrial processes | |
| CN115438726A (en) | Device life and fault type prediction method and system based on digital twin technology | |
| KR20140051678A (en) | Apparatus and method for fault management of smart devices | |
| CN116483054A (en) | Industrial robot running state monitoring and early warning system and method | |
| Mukhopadhyay et al. | Remaining lifetime of degrading systems continuously monitored by degrading sensors | |
| CN117556366B (en) | Data abnormality detection system and method based on data screening | |
| CN118013468B (en) | Method, device, equipment and medium for monitoring health degree of wind turbine generator component | |
| JP7451854B2 (en) | Methods, systems and computer program products for diagnosing energy consumption in industrial environments | |
| CN112345972A (en) | Power failure event-based power distribution network line transformation relation abnormity diagnosis method, device and system | |
| JP2014153957A (en) | Sensor soundness diagnostic device, sensor soundness diagnosis method, plant diagnostic system with sensor soundness diagnostic device and plant diagnostic method by using sensor soundness diagnosis | |
| CN117195451A (en) | Bridge monitoring data restoration method based on graph theory | |
| CN111192163B (en) | Generator reliability medium-short term prediction method based on wind turbine generator operating data | |
| EP3531526A1 (en) | Estimation of electrical grid parameters | |
| KR20200033726A (en) | Data processing method, data processing device, and computer readable recording medium with data processing program thereon | |
| van Dalen et al. | Statistical monitoring of wind turbines |
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 |