Fractional order differential-based wind generating set monitoring data fusion method and equipment
By adopting fractional differential method in the fusion of monitoring data of wind turbines, the problems of monitoring data error and inconsistency are solved, higher data accuracy and signal strength are achieved, and the monitoring efficiency of wind turbines is improved.
Patent Information
- Application Number
- CN202411980834.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-05-23
AI Technical Summary
The prior art has errors and inconsistencies in the fusion of monitoring data of wind turbines, which affects the accuracy of data and the maintenance decisions and operation efficiency of wind turbines.
The data fusion method based on fractional differential is adopted. By obtaining the time series data of the wind turbine unit, calculating the mean value and mean square deviation, constructing the optimal function, and performing fractional differential processing on the monitoring data to obtain the fused monitoring data.
It effectively reduces the error of monitoring data, improves the accuracy of data, enhances signal strength, and improves the accuracy of monitoring data of wind turbines.
Smart Images
Figure CN120030487A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data fusion, and in particular to a method and device for fusion of wind turbine generator monitoring data based on fractional-order differential. Background Art
[0002] With the intensification of the global energy crisis and the improvement of environmental protection awareness, the development and utilization of renewable energy have received unprecedented attention. Wind power generation is one of the most mature and large-scale development potential power generation methods in the field of renewable energy. It has developed rapidly and has become an important support for the transformation of the global energy structure. Wind turbines are usually installed in areas rich in wind energy resources, such as ridges, coastlines and open plains. The environment is complex and changeable. The monitoring and fault diagnosis of their operating status have become the key to ensuring power generation efficiency and equipment safety. With the development of big data and artificial intelligence technology, data fusion technology plays an increasingly important role in the operation monitoring of wind turbines. In order to effectively analyze the operating status of wind turbines, wind turbines are equipped with condition monitoring equipment or systems to form a supervisory control and data acquisition (SCADA) database. However, due to the limitations of sensor accuracy, interference from environmental factors and noise in the data transmission process, the monitoring data often has certain errors and inconsistencies, resulting in inaccurate results, affecting the maintenance decision-making and operation efficiency of wind turbines. The current mainstream data fusion methods have improved the accuracy of data to a certain extent, but there are still some limitations, such as high computing resource requirements, strong dependence on sensor data quality and parameters, and limited generalization ability and real-time performance in different environments. Summary of the invention
[0003] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a wind turbine generator monitoring data fusion method and device based on fractional differential, so as to effectively reduce the monitoring data error and thus improve the accuracy of the monitoring data.
[0004] The purpose of the present invention can be achieved by the following technical solutions:
[0005] A wind turbine generator monitoring data fusion method based on fractional-order differential, the method comprising:
[0006] Obtain time series data for wind turbine status monitoring;
[0007] Sequentially extract multiple groups of continuous time series data of the same number to calculate the average value and mean square error. The average value is used as the monitoring data to form the monitoring data set, and the mean square error is used as the influencing factor to form the influencing factor data set.
[0008] Use monitoring data sets and impact factor data sets to build the optimal function;
[0009] The optimal function is used to perform fractional differential processing on the monitoring data in the monitoring data set to obtain the fused wind turbine monitoring data.
[0010] Furthermore, the time series data includes acceleration, tower inclination, temperature or load collected in chronological order.
[0011] Furthermore, if the mean square error of each set of time series data is less than the threshold, it is determined that the error between the wind turbine monitoring data is small and no subsequent fusion is required.
[0012] Furthermore, the construction process of the optimal function includes:
[0013] Obtain monitoring data and its corresponding impact factors;
[0014] Construct a polynomial function, and use the monitoring data and influencing factors to fit the coefficients of the polynomial function through the least squares method, calculate the mean error of coefficients of different fitting orders, and take the coefficient of the fitting order with the smallest mean error as the optimal coefficient of the polynomial function to construct the optimal function.
[0015] Furthermore, the expression of the polynomial function is:
[0016] F(x)=a 0 +a 1 x+a 2 x 2 +…+a n x n ,
[0017] Among them, F(x) is the polynomial function, n is the fitting order, a n are the coefficients of the polynomial function.
[0018] Furthermore, the formula for the mean error is:
[0019]
[0020] Where F(x) is a polynomial function, F n2 is the mean of the monitoring data, n 2 The number of monitoring data.
[0021] Furthermore, the process of fractional-order differential processing includes:
[0022] Perform fractional differential calculation on the monitoring data to obtain the fractional differential of the monitoring data;
[0023] Calculate the mean of the fractional differential and the mean of the monitoring data, and use the ratio of the two means as the amplification factor of the monitoring signal;
[0024] By dividing each data point in the fractional differential of the monitoring data by the amplification coefficient, the monitoring data after the fractional differential processing is obtained;
[0025] The average value and mean square error of the monitoring data after fractional-order differential processing are calculated, and it is determined whether the mean square error is less than the threshold value. If so, the monitoring data after fractional-order differential processing is output as the fused wind turbine monitoring data. Otherwise, a new optimal function is constructed again according to the monitoring data after fractional-order differential processing, and the above-mentioned fractional-order differential processing is performed on the monitoring data after fractional-order differential processing.
[0026] Furthermore, the formula for calculating the fractional order differential is:
[0027]
[0028] Among them, Γ is the gamma function, v is the differential order, h is the step size, To calculate the number of steps, F(t) is the optimal function with t as the variable, F(t-jh) is the optimal function with t-jh as the variable, D is the differential operator, t is the time point of the monitoring data, a is a constant, when t≤a, the value of the function f(t) is zero, j=0,1,2,…
[0029] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the wind turbine generator monitoring data fusion method based on fractional-order differential are implemented as described above.
[0030] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for fusion of wind turbine generator monitoring data based on fractional-order differentials.
[0031] Compared with the prior art, the beneficial effects of the present invention include: fractional calculus has outstanding advantages in signal processing, especially in terms of signal stability and signal strength enhancement. Fractional calculus can be used to design more efficient filters, improve signal transmission quality, and more accurately describe the physical characteristics of complex systems. At the same time, fractional calculus can describe the memory effect and heritability in the system or process, which cannot be achieved in traditional integer calculus.
[0032] The present invention performs a series of processing on the monitoring data, adopts the mean of the monitoring data as the estimated true value, and corrects the measured estimated true value through fractional differential processing to obtain fused monitoring data, thereby reducing the impact caused by data error fluctuations, effectively reducing the differences between monitoring data and enhancing signal strength, thereby improving the accuracy of wind turbine monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 is a flow chart of the method of the present invention;
[0034] Figure 2 This is an example diagram of least squares fitting of wind turbine acceleration monitoring data according to an embodiment of the present invention;
[0035] Figure 3 This is an example diagram of 0.5-order differential data of wind turbine acceleration monitoring data according to an embodiment of the present invention;
[0036] Figure 4 This is an example diagram of data after 0.5-order differential fusion of wind turbine acceleration monitoring data according to an embodiment of the present invention;
[0037] Figure 5 This is an example diagram of long time series data after 0.57-order differential fusion of wind turbine acceleration monitoring data according to an embodiment of the present invention. DETAILED DESCRIPTION
[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0039] Example 1
[0040] Fractional order differentials have attracted attention due to their advantages in processing memory and non-local characteristic signals. They are suitable for describing nonlinear, non-causal and non-stationary signals, which are similar to the operating environment of wind turbines. In addition, the operation of mechanical components of wind turbines will be affected by historical states, and fractional order differentials can capture memory effects well. This embodiment intends to disclose a method for fusion of wind turbine monitoring data based on fractional order differentials, the method being as follows: Figure 1 As shown, the specific steps include:
[0041] Step S1, obtaining time series data for monitoring the status of a wind turbine generator set. Time series data is data collected at different times and is used to describe the change of a phenomenon over time. This type of data reflects the state or degree of change of a thing or phenomenon over time. The time series data of this embodiment can be acceleration, tower inclination, temperature or load collected in chronological order.
[0042] Step S2, sequentially extracting multiple groups of continuous time series data of the same number to calculate the average value and the mean square error, the average value is used as the monitoring data to form a monitoring data set, and the mean square error is used as the influencing factor to form an influencing factor data set.
[0043] Step S3, compare the mean square error with the threshold. If the mean square error of each set of time series data is less than the threshold, it is determined that the error between the wind turbine monitoring data is small, and no subsequent fusion is required. The data is directly sent to the data acquisition center, otherwise continue with the subsequent steps.
[0044] Step S4, constructing an optimal function using the monitoring data set and the impact factor data set.
[0045] The construction process of the optimal function includes:
[0046] Obtain monitoring data and its corresponding impact factors;
[0047] Construct a polynomial function, and use the monitoring data and influencing factors to fit the coefficients of the polynomial function through the least squares method, calculate the mean error of the coefficients of different fitting orders, and take the coefficient of the fitting order with the smallest mean error as the optimal coefficient of the polynomial function to construct the optimal function.
[0048] The expression of the polynomial function is:
[0049] F(x)=a 0 +a 1 x+a 2 x 2 +…+a n x n ,
[0050] Among them, F(x) is the polynomial function, n is the fitting order, a n are the coefficients of the polynomial function.
[0051] The formula for mean error is:
[0052]
[0053] Where F(x) is a polynomial function, F n2 is the mean of the monitoring data, n 2 The number of monitoring data.
[0054] Step S5, using the optimal function to perform fractional differential processing on the monitoring data in the monitoring data set to obtain fused wind turbine generator set monitoring data.
[0055] The process of fractional differential processing includes:
[0056] Perform fractional differential calculation on the monitoring data to obtain the fractional differential of the monitoring data;
[0057] Calculate the mean of the fractional differential and the mean of the monitoring data, and use the ratio of the two means as the amplification factor of the monitoring signal;
[0058] By dividing each data point in the fractional differential of the monitoring data by the amplification coefficient, the monitoring data after fractional differential processing is obtained and output as the fused wind turbine generator monitoring data.
[0059] The formula for fractional differential calculation is:
[0060]
[0061] Among them, Γ is the gamma function, v is the differential order, h is the step size, To calculate the number of steps, F(t) is the optimal function with t as the variable, F(t-jh) is the optimal function with t-jh as the variable, D is the differential operator, t is the time point of the monitoring data, a is a constant, when t≤a, the value of the function f(t) is zero, j=0,1,2,…
[0062] Step S6, calculate the average value and mean square error of the fused wind turbine monitoring data, and determine whether the mean square error is less than the threshold value. If so, send the fused wind turbine monitoring data to the data acquisition center; otherwise, overwrite the monitoring data in the original monitoring data set and the influencing factors in the original influencing factor data set with the average value and mean square error of the fused wind turbine monitoring data, and return to step S4 for processing again.
[0063] The core idea and processing flow of this embodiment are as follows: considering that the monitoring data is composed of true values and errors, how to eliminate errors is the key to data accuracy. Under actual working conditions, the true value is difficult to measure directly and is an ideal value. The present invention adopts the mean as the estimated true value, which is associated with the data error, and a mathematical model of the estimated true value and the data error can be established. Generally speaking, the true value is a constant value under a certain monitoring state. If the data error fluctuates greatly, it will cause the measured value to shift up and down and deviate from the true value. Therefore, the impact caused by the fluctuation of the data error should be reduced, and the measured data should be corrected, so that the measured data can be closer to the true value. The differential of the estimated true value and the data error can effectively reflect the impact of the data error. The estimated true value of the measurement can be corrected by the fractional differential method.
[0064] Below, we simulate the actual scenario and give an application example:
[0065] The acceleration time series data of a wind farm in a certain period of time is calculated every n 1 =10 data are grouped together, the average value is calculated as the monitoring data F, and the standard deviation S of each group of data is calculated as the influencing factor x.
[0066] Take n 2 =10, get 10 sets of monitoring data F 10 , with the polynomial F(x)=a 0 +a1 x+a 2 x 2 +…+a n x n Fitting function, n<10. The least square method is used to obtain the parameters of the fitting function. The fitting order is from the lowest to the highest, and the mean error of each order is calculated respectively. The order with the smallest mean error is the optimal function F(x). The data fitting results are as follows: Figure 2 .
[0067] The fractional order differential model defined by Grunwald-Letnikov is used for calculation. The differential order v = 0.5, the step size h = 0.001. The fractional order differential calculation is performed for the monitoring data set F. The calculation results are shown in Figure 3 As shown, the fractional differential data set F 0.5 Compared with the original data set F, the increase is significantly increased. Calculate the amplification factor F 0.5 Each data in is divided by K, and the acceleration monitoring data f after 0.5 order differential processing is obtained. 0.5 The final result of the 0.5-order differential processing is as follows Figure 4 It can be seen that the monitoring data processed by the fractional-order differential operator are evenly distributed around their average value, and the variability between the monitoring data is significantly reduced, indicating that the fractional-order differential operator has significant advantages in fusing data with large differences.
[0068] In order to achieve the optimal differential order, within the range of (0.1, 0.9), traverse with a step size of 0.01, follow the above calculation process, take 10,000 acceleration monitoring data as samples, obtain the differential order corresponding to the maximum value of the amplification factor, and the optimal differential order is 0.57. The results are as follows Figure 5 As shown in Figure 2, it can be seen that the volatility of the processed monitoring data is greatly reduced, distributed around the mean, and the standard deviation is reduced by 71%.
[0069] Furthermore, in another embodiment, a fractional-order differential model with other definitions can be used for calculation.
[0070] Example 2
[0071] Based on Example 1, this embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the aforementioned wind turbine monitoring data fusion method based on fractional-order differentials.
[0072] At the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the above-mentioned wind turbine monitoring data fusion method based on fractional differentials. Of course, in addition to the software implementation, the present invention does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.
[0073] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0074] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0075] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. A wind turbine monitoring data fusion method based on fractional differential, characterized in that: The method comprises: Obtain time series data for wind turbine status monitoring; Sequentially extract multiple groups of continuous time series data of the same number to calculate the average value and mean square error. The average value is used as the monitoring data to form the monitoring data set, and the mean square error is used as the influencing factor to form the influencing factor data set. Use monitoring data sets and impact factor data sets to build the optimal function; The optimal function is used to perform fractional differential processing on the monitoring data in the monitoring data set to obtain the fused wind turbine monitoring data.
2. According to claim 1, a method for wind turbine monitoring data fusion based on fractional differential is characterized in that: The time series data includes acceleration, tower inclination, temperature or load collected in chronological order.
3. The method for wind turbine monitoring data fusion based on fractional differential according to claim 1 is characterized in that: If the mean square error of each set of time series data is less than the threshold, it is determined that the error between the wind turbine monitoring data is small and no subsequent fusion is required.
4. The method for wind turbine monitoring data fusion based on fractional differential according to claim 1 is characterized in that: The construction process of the optimal function includes: Obtain monitoring data and its corresponding impact factors; Construct a polynomial function, and use the monitoring data and influencing factors to fit the coefficients of the polynomial function through the least squares method, calculate the mean error of the coefficients of different fitting orders, and take the coefficient of the fitting order with the smallest mean error as the optimal coefficient of the polynomial function to construct the optimal function.
5. The method for wind turbine monitoring data fusion based on fractional differential according to claim 4 is characterized in that: The expression of the polynomial function is: F(x)=a0+a1x+a2x 2 +…+a n x n , Among them, F(x) is the polynomial function, n is the fitting order, a n are the coefficients of the polynomial function.
6. The method for wind turbine monitoring data fusion based on fractional differential according to claim 4 is characterized in that: The formula for the mean error is: Where F(x) is a polynomial function, F n2 is the mean of the monitoring data, and n2 is the number of monitoring data.
7. The method for wind turbine monitoring data fusion based on fractional differential according to claim 1 is characterized in that: The process of fractional differential processing includes: Perform fractional differential calculation on the monitoring data to obtain the fractional differential of the monitoring data; Calculate the mean of the fractional differential and the mean of the monitoring data, and use the ratio of the two means as the amplification factor of the monitoring signal; By dividing each data point in the fractional differential of the monitoring data by the amplification coefficient, the monitoring data after the fractional differential processing is obtained; The average value and mean square error of the monitoring data after fractional-order differential processing are calculated, and it is determined whether the mean square error is less than the threshold value. If so, the monitoring data after fractional-order differential processing is output as the fused wind turbine monitoring data. Otherwise, a new optimal function is constructed again according to the monitoring data after fractional-order differential processing, and the above-mentioned fractional-order differential processing is performed on the monitoring data after fractional-order differential processing.
8. The method for wind turbine generator monitoring data fusion based on fractional differential according to claim 7, characterized in that: The formula for fractional differential calculation is: Among them, Γ is the gamma function, v is the differential order, h is the step size, To calculate the number of steps, F(t) is the optimal function with t as the variable, F(t-jh) is the optimal function with t-jh as the variable, D is the differential operator, t is the time point of the monitoring data, a is a constant, when t≤a, the value of the function f(t) is zero, j=0,1,2,… 9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the wind turbine generator monitoring data fusion method based on fractional-order differentials as described in any one of claims 1 to 8 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the wind turbine generator monitoring data fusion method based on fractional-order differentials as described in any one of claims 1 to 8 are implemented.