JP2016035208A - Wind power generating facility stress estimation apparatus, wind power generating facility stress estimation method, and wind power generating system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a wind power generating facility stress estimation apparatus capable of measuring a stress of a wind turbine or monitoring the stress without providing a stress detection sensor beyond necessity.SOLUTION: A wind power generating facility stress estimation apparatus comprises: a similar wind condition selection unit 10 for selecting a similar wind condition of another wind turbine having the similar wind condition from input wind condition data; a stress value database 20 for storing stress values of a plurality of wind turbines; an estimated stress value calculation unit 30 for estimating a stress value of a stress-unmeasured wind turbine from the stress values stored in the stress value database 20 corresponding to the selected wind turbine; and a stress value database update unit 40 for updating the stress values of the plurality of wind turbines stored in the stress value database 20 on the basis of the estimated stress value S20 calculated by the estimated stress value calculation unit 30. The estimated stress value calculation unit 30 estimates the stress value of the stress-unmeasured wind turbine on the basis of a plurality of similarities of wind conditions selected by the similar wind condition selection unit 10 and a plurality of stress values extracted from the stress value database 20.SELECTED DRAWING: Figure 1

Description

  The present invention relates to wind power generation equipment, and more particularly to structural strength diagnosis of wind power generation equipment.

  From the viewpoint of securing stable energy resources and preventing global warming, great expectations are placed on the introduction and expansion of renewable energy such as solar power generation and wind power generation. In the design of wind power generation systems, to improve power generation efficiency, increase the size of wind power generation facilities, streamline construction methods, shorten construction periods, simplify maintenance, repairs and inspections, and reduce power generation costs and make it easier to establish a power generation business Various studies are underway.

  With the increase in the size of wind power generation facilities, in the wind power generation system of several MW scale, the blade length of the windmill and the height of the tower reach several tens of meters. For this reason, the force (moment) applied to the blade and the tower greatly changes due to the change in the wind speed and the wind direction received by the windmill, and a large stress (strain) is generated in the blade and the tower. Fatigue accumulates in the blade and tower due to this stress, which can lead to accidents such as blade breakage and tower collapse.

  Therefore, efforts are being made to reduce the loss cost and improve the power generation profit by monitoring the state of the wind power generation equipment and implementing optimal control and feedback to the design.

  As a background art in this technical field, for example, there is a technique such as Patent Document 1. Patent Document 1 discloses a stress analysis device for a wind turbine structure that creates stress time series data to be analyzed from time series data of loads.

  Patent Document 2 discloses a method for setting an evaluation index for durability strength during wind turbine operation, in which stress time series data to be analyzed is created from load time series data.

  Patent Document 3 discloses a design method for a floating wind turbine generator that obtains stress to be analyzed from time-series data of wind load and wave load.

  Patent Document 4 discloses a wind power generation system that predicts a wind condition based on time-series wind condition data and controls a wind turbine.

  Further, Patent Document 5 discloses a power generation amount prediction method that predicts fluctuations in wind speed based on time-series wind condition data and the like, and performs output control of a generator of a wind power generation facility.

JP 2010-79585 A WO2010 / 038305 JP-A-2005-240785 JP 2008-64081 A JP 2013-222423 A

  As mentioned above, in wind power generation equipment, it is possible to efficiently monitor the state of the windmill, especially the stress (strain) generated in the windmill, and feed back to the optimum operation control, windmill blade and tower strength design. It is an important issue in predicting the life of a windmill, preventing accidents such as blade breakage and tower collapse, and supplying power stably.

  However, in the structural strength of the machine, there is a large discrepancy between the test data and the actually measured data on site, and there are many cases where damage occurs earlier than expected. Moreover, when measuring the structural strength of a windmill, it is necessary to install a large number of sensors in each windmill, but there are problems of sensor cost and securing the installation location, and it is difficult to measure multiple windmills.

  In the stress analysis device for a wind turbine structure disclosed in Patent Document 1, load time series data at a predetermined load observation point set in the wind turbine structure is created, and the wind turbine structure is set based on the load time series data. Stress time series data at at least one analysis target location is calculated by the stress analysis means.

  However, in this method, in the stress analysis means, it is necessary to create the load time series data of the analysis target location from the load data of the load observation location using the load conversion table associated with the analysis target location. It is practically impossible to prepare a load conversion table for combinations of load observation locations and all analysis target locations, and the analysis target locations are actually limited.

  In the evaluation index setting method of Patent Document 2, load time-series data at a predetermined load observation point set in the wind turbine structure is created, stress time-series data is obtained based on the load time-series data, and stress time-series data is obtained. When the stress based on is applied to the evaluation target part over the compensation operation period, the minimum fracture toughness value is determined so that the evaluation target part does not cause fragile fracture.

  However, in this method, the load observation point and the evaluation target point are the same, and therefore it is not possible to reduce the number of sensors when it is desired to evaluate a large number of points.

  In the design method of the floating wind power generator of Patent Document 3, the stress amplitude probability distribution is obtained from the statistical value of the stress due to the waves, the fatigue damage degree due to the waves is obtained from the probability density distribution of the stress amplitudes, and the stress due to the wind is calculated. Obtain the degree of fatigue damage due to wind from the time series. With the combined fatigue damage level that is the sum of the fatigue damage level due to waves and wind, it is possible to accurately estimate the effects of repeated loads due to wind and waves and reflect them in the fatigue strength design of floating wind turbine generators. it can.

  However, in this method, as in Patent Document 2, the load observation location and the evaluation target location are the same, and therefore it is not possible to reduce the number of sensors when evaluating a large number of locations.

  In the wind power generation system of Patent Document 4, a predictive control amount estimation mechanism that maximizes the profit related to the power generation amount and cost is introduced using the predicted wind condition as a control input, and can be obtained from a single or a plurality of wind condition observation mechanisms. Precise wind condition prediction is performed using optimal dynamical system reconstruction from the three-dimensional wind condition observation series. As a result, appropriate space-time reconstruction of the regional wind dynamics system that makes the best use of the spatio-temporal information of the wind conditions, high-precision prediction, and control without waste are made, and highly efficient power generation becomes possible.

  However, with this method, it is necessary to reduce the prediction error in order to predict the wind conditions, but it is difficult to accurately perform weather forecasts. Further, no mention is made regarding the structural strength of the wind turbine.

  In the power generation amount prediction method of Patent Document 5, the power generation amount prediction of an order wind power generation facility that is several tens of minutes later as a margin for performing output control of the power system side generator is realized. The present invention stores past wind condition time series data and past weather time series data, and extracts past wind condition time series data similar to the current wind condition / meteorological time series data. In addition, wind conditions are predicted to estimate the amount of power generation.

  However, in this method, as in Patent Document 4, it is necessary to reduce the prediction error in order to predict the wind conditions, but it is difficult to accurately perform weather forecasting. Similarly, the structural strength of the wind turbine is not mentioned.

  Accordingly, an object of the present invention is to provide a stress estimation device for a wind turbine generator capable of measuring or monitoring the stress of a wind turbine without providing a stress detection sensor more than necessary.

  Another object of the present invention is to provide a stress estimation method for wind power generation equipment capable of measuring or monitoring stress of a wind turbine without providing a stress detection sensor more than necessary.

  Another object of the present invention is to provide a wind power generation system capable of measuring or monitoring stress of a wind turbine without providing a stress detection sensor more than necessary.

  In order to solve the above problems, the present invention selects another wind turbine having a wind condition similar to the wind condition data based on the measured wind condition data, and determines a wind turbine whose stress has not been measured from the stress value of the selected wind turbine. A stress estimation device for a wind power generation facility that estimates a stress value of the wind power generation facility, wherein the stress estimation device for the wind power generation facility selects a wind condition of another windmill having a similar wind condition from input wind condition data. A state selection unit, a stress value database that accumulates stress values of a plurality of wind turbines, and an estimated stress value calculation unit that estimates a stress value of a wind turbine that has not been measured from the stress values of the stress value database corresponding to the selected wind turbine A stress value database update unit that updates the stress values of a plurality of wind turbines accumulated in the stress value database based on the stress estimated value calculated by the estimated stress value calculation unit, and the estimated stress value calculation unit Before Based on a plurality of stress values extracted from the selected plurality of wind conditions similarity and the stress value database in a similar wind conditions selecting unit, and estimates the stress values of the wind turbine of the stress unmeasured.

  Further, the present invention selects other wind turbines whose wind conditions are similar to the wind condition data based on the measured wind condition data, and estimates the stress value of the wind turbine whose stress has not been measured from the stress value of the selected wind turbine. A stress estimation method for a wind power generation facility, wherein a wind classification and a wind similarity of wind conditions similar to the measured wind data are selected, and stress values corresponding to the selected wind classification are stored in a stress value database. And the stress value of the wind turbine whose stress has not been measured is estimated based on the stress value extracted from the stress value database and the wind condition similarity.

  Further, the present invention is a wind power generation system for estimating a stress value of a windmill whose stress has not been measured based on measured wind condition data, wherein the wind power generation system stores stress values of a plurality of windmills. The stress values of the wind turbines that are not measured from the stress values of the plurality of wind turbines extracted from the stress value storage device, and the stress values of the wind turbines and other wind turbines whose wind conditions are similar to the measured wind condition data And a stress estimation device for estimating a value.

  ADVANTAGE OF THE INVENTION According to this invention, the stress estimation apparatus of the wind power generation equipment which can measure the stress of a windmill or can monitor a stress can be implement | achieved, without providing a stress detection sensor more than necessary.

  Further, according to the present invention, it is possible to realize a stress estimation method for a wind power generation facility capable of measuring or monitoring the stress of a windmill without providing a stress detection sensor more than necessary.

  Further, according to the present invention, it is possible to realize a wind power generation system capable of measuring stress or monitoring stress of a windmill without providing a stress detection sensor more than necessary.

  In addition, according to the present invention, it is possible to predict the life of the windmill without providing an unnecessarily stress detection sensor, prevent accidents such as blade breakage and tower collapse, and enable stable power supply. .

Moreover, according to this invention, the stress of a windmill can be estimated before construction of a wind power generation facility.
Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is a figure which shows the structure of the wind power generation system which concerns on one Embodiment of this invention. It is a figure which shows the wind condition classification | category in the similar wind condition selection part of the wind power generation system which concerns on one Embodiment of this invention. It is a flowchart which shows the stress estimation method of the wind power generation system which concerns on one Embodiment of this invention. It is a figure showing the whole wind power generation system outline concerning one embodiment of the present invention.

  Embodiments of the present invention will be described with reference to the drawings. In each drawing and each embodiment, the same or similar components are denoted by the same reference numerals, and description thereof is omitted.

  FIG. 1 is a diagram showing a configuration of a wind power generation system according to an embodiment of the present invention. The wind power generation system 100 shown in FIG. 1 receives, for example, a wind condition measurement value S10 that is data from a sensor installed in a wind turbine to be diagnosed, and the wind condition of the wind turbine to be diagnosed based on the wind condition measurement value S10. Class (for example, wind speed is strong, wind direction is southeast, etc.), similar wind condition selection unit 10 for calculating similar wind condition data and its similarity, and wind condition class selected by similar wind condition selection unit 10 A stress value DB 20 for outputting a stress value corresponding to a wind condition class input from a database (DB) storing a plurality of wind condition classes and stress values corresponding to each, and a similar wind condition selection unit The estimated stress that calculates the unmeasured stress estimated value by calculating the similarity of the wind class selected in 10 and the stress value output from the stress value DB 20 and calculating a plurality of wind similarity and stress values. Calculation unit 30 and unmeasured stress estimation A constant value S20 is input, compared with an appropriate stress value stored in the stress value DB 20, a stress value DB update unit 40 for updating the stress value DB 20 is provided, and an unmeasured stress estimated value S20 is output.

  The wind condition measurement value S10 is signal data for sensing the state of the wind, and includes, for example, meteorological data such as wind direction, wind speed, temperature, humidity, rainfall, and weather, and terrain data.

  In the similar wind condition selection unit 10, the wind condition measurement value S10 is input, and a plurality of wind condition classes prepared in advance are compared with the wind condition measurement value S10 of the wind power generation system whose stress is to be estimated, and the similarity is calculated. Select multiple wind conditions classes with high similarity (select labels). The wind condition class is configured by combining a plurality of arbitrary types of signals from several kinds of signal data constituting the wind condition measurement value S10.

  FIG. 2 is a table showing an example of wind condition class classification in the similar wind condition selecting unit 10 of FIG. Basically, since the description has already been repeated, detailed description will be omitted. However, as shown in FIG. 2, the wind speed is combined with the wind direction and the wind speed in the signal data for sensing the wind state. 24 classes are created by combining 3 levels (weak, medium, strong) and 8 wind directions (north, northeast, east, southeast, south, southwest, west, northwest), and a label is assigned to each class.

  The similarity of wind conditions is machine learning that learns regularity such as frequent pattern extraction and class classification, using the signal data set that constitutes the wind condition class in the wind condition measurement value S10, that is, principal component analysis, clustering, support Using a method such as a vector machine, the similarity to an arbitrary wind condition class is calculated. This is calculated by the number of selected wind class, and a plurality of pairs of wind class having high similarity and its similarity are calculated. At that time, the measured wind condition measurement values of other wind turbines stored in the stress value DB 20 are learned in advance as learning data.

  In addition to the above, the calculation of the similarity of the wind conditions includes, for example, the principal component analysis of the wind condition data input to the similar wind condition selection unit 10 and the deviation (variation) from the average value of each parameter of the wind condition data. It can also be calculated using.

  Alternatively, a clustering method based on the distance between each wind turbine group in a plurality of wind turbine groups, comparison of maximum values for each wind turbine group in a plurality of wind turbine groups, comparison of average values, comparison of variations, variance, deviation, etc. good.

  In the stress value DB 20, the measured stress values of the windmill are accumulated. Each data includes a measurement date, a wind condition measurement value (meteorological data) on the measurement date, a measurement time, a measurement site, a stress value, and the like, and is labeled with a wind condition class. There are as many labels as there are classes classified by the similar wind condition selection unit 10. According to the label of the wind condition class selected by the similar wind condition selection unit 10, an appropriate stress value is output to the estimated stress calculation unit 30.

  The estimated stress calculation unit 30 receives the similarity of the wind class selected by the similar wind condition selection unit 10 and the stress value output from the stress value DB 20 and inputs unmeasured stress from the plurality of wind similarity and stress values. Estimated value S20 is calculated.

  The calculation method can use, for example, a product-sum operation of a plurality of wind condition similarities and stress values as shown in Equation 1.

Unmeasured stress estimated value = Σ (wind condition similarity A × stress value A) + (wind condition similarity B × stress value B) + ...: Formula 1
Alternatively, as shown in Equation 2, the stress with similar wind conditions can be obtained from the product of the measured wind turbine stress value A and the reciprocal of the wind condition similarity A.

Unmeasured stress estimated value = stress value A × (1 / wind condition similarity A): Formula 2
Moreover, a predetermined threshold value may be provided for each of a plurality of wind state similarities and stress values, and an unmeasured stress estimated value may be calculated using wind state similarity and stress values that are equal to or greater than the threshold values.

  Alternatively, when a certain wind condition similarity is larger than other wind condition similarities, a stress value corresponding to the wind condition similarity may be regarded as an unmeasured stress estimated value.

  The stress value DB updating unit 40 updates the stress value DB 20 based on the unmeasured stress estimated value S20 calculated by the estimated stress value calculating unit 30.

  When the wind state similarity of the unmeasured stress estimated value S20 is 1 for a single wind state label in the stress value DB 20, that is, when it is a perfect match, the stress value DB 20 is not updated, but there are a plurality of wind state similarities, If the similarity is low, the data does not exist in the stress value DB 20, and therefore the unmeasured stress estimated value S20 is accumulated and the stress value DB 20 is updated.

  FIG. 3 is a flowchart illustrating an example of a stress estimation method for a wind power generation system according to an embodiment of the present invention. Basically, since it has already been repeated, detailed description is omitted. However, as shown in FIG. 3, when the stress estimation is started, the wind condition measurement value S10 is input in the similar wind condition selection step F1. Then, the wind conditions of the wind condition measurement value S10 are classified, the wind condition label S30 and the wind condition similarity S40 are obtained, and the process proceeds to the stress value DB access step F2.

  In the stress value DB access step F2, the stress value DB 20 is accessed using the wind condition label S30 obtained in the similar wind condition selection step F1 as a key, and an appropriate stress value is output to the estimated stress value calculation step F3. To do.

  In the estimated stress value calculation step F3, the stress value output from the stress value DB access step F2 and the wind condition similarity S40 output from the similar wind condition selection step F1 are input, and the unmeasured stress estimated value S20 is, for example, It is calculated by the product-sum operation as shown in Equation 1 above. The unmeasured stress estimated value S20 is output and fed back to the stress value DB update step F4.

  The stress value DB update step F4 receives the unmeasured stress estimated value S20 as an input, updates the stress value stored in the stress value DB 20, and ends.

  FIG. 4 is a diagram showing an overall outline of a wind power generation system according to an embodiment of the present invention.

  The windmill 200 includes a tower E10 or a blade E20 that is a stress estimation target device, and a stress estimation device 300 that diagnoses a state based on the wind condition measurement value S10. The stress estimation apparatus 300 is the same as the configuration of the wind power generation system 100 described in the embodiments so far except for the stress value DB 20 and the stress value DB update unit 40, and the stress values of unmeasured locations in the tower and blades. Is estimated.

  The stress value storage device 400 including the stress value DB 20 and the stress value DB update unit 40 is placed in a facility connected to a plurality of windmills such as a wind farm monitoring building 500 through a network.

  As described above, according to the present invention, it is possible to perform stress measurement or stress monitoring of a wind turbine without providing a stress detection sensor more than necessary.

  In addition, it is possible to predict the life of the windmill without providing a stress detection sensor more than necessary, prevent accidents such as blade breakage and tower collapse, and supply power stably.

Moreover, the stress of a windmill can be estimated before construction of a wind power generation facility.
In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

  In addition, each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them, for example, with an integrated circuit.

  Further, each of the above-described configurations and functions may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as programs, tables, and files for realizing each function is recorded on a recording device such as a memory, a hard disk, or an SSD (Solid-State-Drive), or a recording medium such as an IC card, a memory card, or a DVD. You can also.

  In addition, the control lines and information lines are those that are considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

  DESCRIPTION OF SYMBOLS 10 ... Similar wind condition selection part, 20 ... Stress value DB, 30 ... Estimated stress value calculation part, 40 ... Stress value DB update part, 100 ... Wind power generation system, 200 ... Windmill, 300 ... Stress estimation apparatus, 400 ... Stress value Accumulator, 500 ... Wind farm monitoring ridge, E10 ... Tower, E20 ... Blade, F1 ... Similar wind condition selection step, F2 ... Stress value DB access step, F3 ... Estimated stress value calculation step, F4 ... Stress value DB update step, S10: wind condition measurement value, S20: unmeasured stress estimation value, S30: wind condition label, S40: wind condition similarity.

Claims (15)

Based on the measured wind condition data, select another windmill with a similar wind condition to the wind condition data, and estimate the stress value of the unmeasured wind turbine from the stress value of the selected wind turbine. A device,
The stress estimation device of the wind power generation facility includes a similar wind condition selection unit that selects a wind condition of another windmill having a similar wind condition from the input wind condition data,
A stress value database for accumulating stress values of a plurality of wind turbines;
An estimated stress value calculation unit that estimates a stress value of a windmill whose stress is not measured from a stress value of the stress value database corresponding to the selected windmill;
A stress value database update unit that updates the stress values of a plurality of wind turbines accumulated in the stress value database, based on the estimated stress value calculated by the estimated stress value calculation unit,
The estimated stress value calculation unit estimates a stress value of a windmill whose stress has not been measured based on a plurality of wind state similarities selected by the similar wind state selection unit and a plurality of stress values extracted from the stress value database. A stress estimation apparatus for wind power generation equipment.   The wind condition data input to the similar wind condition selection unit uses at least two or more meteorological data among wind direction, wind speed, temperature, humidity, rainfall, and weather meteorological data. The stress estimation apparatus of the described wind power generation equipment.   The similar wind condition selection unit includes a principal component analysis of wind condition data input to the similar wind condition selection unit, a deviation from an average value of each parameter of the wind condition data input to the similar wind condition selection unit, a plurality of The other wind turbines having similar wind conditions are selected by any one of a clustering method based on a distance between the wind turbine groups in the wind turbine group, frequent pattern extraction, or machine learning for learning regularity of class classification. Item 3. The stress estimation device for wind power generation equipment according to item 1 or 2.   The estimated stress value calculation unit estimates the stress of the unmeasured wind turbine by performing a product-sum operation on the plurality of wind condition similarities and the plurality of stress values. The stress estimation apparatus of the described wind power generation equipment.   The stress estimation device for wind power generation equipment according to any one of claims 1 to 4, wherein the part for estimating the stress value in the windmill whose stress has not been measured is a tower or a blade of the windmill. Based on the measured wind condition data, select another windmill with a similar wind condition to the wind condition data, and estimate the stress value of the unmeasured wind turbine from the stress value of the selected wind turbine. A method,
Select a wind classification and wind similarity of wind conditions similar to the measured wind data,
Extracting stress values corresponding to the selected wind classification from a stress value database;
A stress estimation method for wind power generation equipment, characterized by estimating a stress value of a wind turbine whose stress has not been measured based on a stress value extracted from the stress value database and the wind condition similarity.   The wind power generation equipment stress estimation method according to claim 6, wherein the wind condition data uses at least two or more meteorological data among wind direction, wind speed, temperature, humidity, rainfall, and weather meteorological data. .   Learn the principal component analysis of the wind data, the deviation from the average value of each parameter of the wind data, the clustering method based on the distance between each wind turbine group in multiple wind turbine groups, the regularity of frequent pattern extraction or class classification The stress estimation method for wind power generation equipment according to claim 6 or 7, wherein a similar wind condition is selected by any of machine learning to be performed.   The stress estimation method for a wind power generation facility according to claim 6 or 7, wherein the stress of the windmill without stress measurement is estimated by performing a product-sum operation on a plurality of wind state similarities and a plurality of stress values.   The stress estimation method for wind power generation equipment according to any one of claims 6 to 9, wherein the part for estimating the stress value in the windmill whose stress has not been measured is a tower or a blade of the windmill. A wind power generation system that estimates the stress value of an unmeasured wind turbine based on measured wind condition data,
The wind power generation system includes a stress value accumulation device that accumulates stress values of a plurality of wind turbines,
The stress values of a plurality of other wind turbines whose wind conditions are similar to the measured wind condition data are extracted from the stress value storage device, and the stress values of the unmeasured wind turbine are estimated from the stress values of the extracted wind turbines A wind power generation system comprising:   The wind power generation system according to claim 11, wherein the wind condition data uses at least two or more weather data among wind direction, wind speed, temperature, humidity, rainfall, and weather meteorological data.   In the stress estimation device, a principal component analysis of the wind data, a deviation from an average value of each parameter of the wind data, a clustering method based on a distance between wind turbine groups in a plurality of wind turbine groups, frequent pattern extraction or class The wind power generation according to claim 11 or 12, wherein stress values of a plurality of other wind turbines having similar wind conditions are extracted from the stress value accumulating device by any of machine learning for learning the regularity of classification. system.   13. The wind power generation according to claim 11, wherein the stress estimation device estimates the stress of the windmill not measuring the stress by performing a product-sum operation on a plurality of wind condition similarities and a plurality of stress values. system.   The wind power generation system according to any one of claims 11 to 14, wherein a part for estimating a stress value in the unmeasured windmill is a tower or a blade of the windmill.

Images