A universal vibration monitoring and collecting performance test method for doubly-fed and direct-driven wind turbines
By using integrated operating condition setup and hierarchical synchronous sampling, the problem of model compatibility for vibration acquisition and testing of doubly fed and direct-drive wind turbines was solved, realizing a standardized testing process with high precision, low false positive rate and low loss, which can meet the needs of batch testing.
Patent Information
- Application Number
- CN202611492206.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-23
- Publication Date
- 2026-11-06
AI Technical Summary
In the existing technology, vibration acquisition and testing of doubly-fed fans and direct-drive fans have problems such as incompatibility between models, insufficient accuracy due to lack of quantization correction, time-consuming step-by-step testing, high misjudgment rate in single-test, and large losses due to frequent start-ups and shutdowns of equipment.
The test method adopts integrated working condition setup, layered synchronous vibration sampling, model quantization deviation compensation, multi-level progressive performance discrimination, constant working condition closed-loop re-inspection, and environmental adaptive threshold correction to realize a unified test process for doubly-fed and direct-drive wind turbines. By differentiating synchronous sampling, quantization correction, and multiple re-inspections, the differences between models and environmental interference are eliminated, thereby improving test accuracy and reliability.
It has achieved unified testing standards and reliability for both doubly-fed and direct-drive wind turbines, improved testing accuracy, shortened testing time, reduced misjudgment rate and equipment wear, and adapted to the needs of batch standardized testing.
Smart Images
Figure CN123045518A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wind power equipment monitoring performance testing technology, and in particular to a universal vibration monitoring and acquisition performance testing method for doubly fed and direct-drive wind turbines. Background Technology
[0002] Doubly fed induction generators (DFIGs) and direct-drive wind turbines are the mainstream wind power models. Their vibration characteristics differ significantly; DFIGs primarily exhibit low-frequency fluctuating vibrations, while direct-drive wind turbines primarily exhibit high-frequency steady-state micro-vibrations. Existing general vibration acquisition and performance testing technologies suffer from four major shortcomings: a lack of standardized quantitative algorithms, and poor overall testing standardization and reliability.
[0003] Existing technologies employ a step-by-step testing mode for different models, lacking an integrated working condition adaptation and synchronous sampling mechanism. This results in redundant and time-consuming testing procedures. Furthermore, the lack of a quantitative compensation model, relying solely on manual experience for correction, leads to low testing accuracy and poor data consistency. The use of only single-data analysis fails to filter out random environmental interference, resulting in a high false positive rate. Frequent switching of working conditions and equipment start-ups and shutdowns exacerbate hardware wear and tear on the tested acquisition units, leading to high overall testing costs and making it unsuitable for batch standardized testing needs. Summary of the Invention
[0004] The purpose of this invention is to provide a universal vibration monitoring and acquisition performance testing method for both doubly-fed and direct-drive wind turbines, solving the technical problems of existing wind turbine vibration acquisition and testing equipment such as incompatible models, insufficient accuracy due to lack of quantization correction, lengthy step-by-step testing, high single-judgment misjudgment rate, and high losses due to frequent equipment start-ups and shutdowns.
[0005] This invention provides a universal vibration monitoring and acquisition performance testing method for both doubly-fed and direct-drive wind turbines, comprising the following steps: Step S1: Integrated working condition setup; Construct a general simulation test condition that simultaneously covers the variable speed fluctuation vibration of doubly fed wind turbines and the steady-state high-frequency vibration characteristics of direct-drive wind turbines, and connect the vibration acquisition unit under test to the standardized test link; Step S2: Layered synchronous vibration sampling; Based on the vibration differences between the two types of wind turbines, a differentiated synchronous sampling strategy is adopted to collect multi-dimensional original vibration signals of the two types of wind turbines in parallel and construct the original vibration dataset; Step S3: Model Quantization Deviation Compensation; Call the preset model deviation compensation mathematical model to quantize and correct the original vibration dataset, eliminate the data offset caused by the inherent operating condition differences between the two types of wind turbines, and obtain the corrected vibration dataset. Step S4: Multi-level progressive performance judgment; A two-level progressive verification logic is adopted. Based on the calibrated vibration dataset, the signal integrity verification, acquisition accuracy and response performance judgment are completed in sequence, and the preliminary test results are output. Step S5: Closed-loop retest under constant operating conditions; keep the test conditions and environmental parameters unchanged, and conduct multiple retests on the initial abnormal samples to distinguish between random interference pseudo-anomalies and real performance defects. Step S6: Environmental Adaptive Threshold Correction; Collect real-time environmental parameters, dynamically update the judgment threshold through the environmental correction model, output the final standardized test result, and complete the closed-loop test.
[0006] Preferably, in step S2, the differentiated sampling strategy for layered synchronous vibration sampling is as follows: a continuous steady-state sampling mode is used for the low-frequency fluctuation vibration signal of the doubly fed fan, and an intermittent encrypted sampling mode is used for the high-frequency transient micro-vibration signal of the direct-drive fan. The two types of sampling are aligned in time, correspond in measurement points, and are executed synchronously in parallel.
[0007] Preferably, in step S3, the mathematical model calculation formula for model deviation compensation is as follows: ; in, X This represents the original vibration amplitude. Y The corrected vibration amplitude; K This refers to the model compensation coefficient; This is the baseline threshold for the model. Corresponding to doubly fed wind turbines; Corresponding to direct drive fans.
[0008] Preferably, in step S4, the two-level progressive performance discrimination logic includes: the first level verifies the continuity, integrity, and absence of missing information in the vibration signal; the second level compares the consistency between the correction data and the standard sample data, and simultaneously verifies the signal response delay index.
[0009] Preferably, in step S5, the closed-loop retesting mechanism is as follows: for the initially marked abnormal samples, all test parameters are kept constant, and three consecutive closed-loop retests are performed. If the results of the three tests are consistent and deviate, it is determined to be a true performance abnormality.
[0010] Preferably, in step S6, the environmental adaptive correction is implemented based on an environmental correction model, the model formula of which is shown below: ; in, This is the threshold correction coefficient; Real-time temperature; Reference temperature; Electromagnetic interference intensity; and The adjustment weights are fixed.
[0011] Preferably, in step S2, the two sampling modes share the same signal acquisition channel and time axis, without the need to switch hardware links, ensuring that the data of the two models can be compared horizontally.
[0012] Preferably, in step S3, the model deviation compensation only performs algorithm calculations on the post-dataset and does not modify the original acquisition logic and operating parameters of the unit under test.
[0013] Preferably, in step S5, after the re-inspection is completed, all valid samples are integrated to generate a standardized test report with dual model classification labels, forming a complete test closed loop.
[0014] Therefore, the present invention adopts the above-mentioned universal vibration monitoring and acquisition performance test method for doubly fed and direct-drive wind turbines, which solves the technical problems of existing wind turbine vibration acquisition and testing equipment model mismatch, insufficient accuracy due to lack of quantization correction, lengthy step-by-step testing time, high single-judgment misjudgment rate, and large losses due to frequent equipment start-up and shutdown.
[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of the overall process of a universal vibration monitoring and acquisition performance testing method for doubly-fed and direct-drive wind turbines according to the present invention; Figure 2 This is a comparative bar chart of Example 1, Comparative Example 1, and Comparative Example 2 in the universal vibration monitoring and acquisition performance test method for doubly-fed and direct-drive wind turbines of the present invention. Detailed Implementation
[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0018] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.
[0019] The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the term encompasses the elements or objects listed after the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0020] Example 1, Comparative Example 1, and Comparative Example 2 used identical test hardware, test environment, test samples, and test duration benchmarks. The ambient benchmark temperature was 25℃, and there was no strong electromagnetic interference. The same general-purpose wind turbine vibration acquisition unit was selected, and the single benchmark test duration was 50 minutes. The simulated operating parameters of doubly-fed and direct-drive wind turbines were unified, and all variables were limited to differences in test methods and algorithms.
[0021] Example 1 like Figures 1-2 As shown, the present invention provides a universal vibration monitoring and acquisition performance testing method for both doubly-fed and direct-drive wind turbines, comprising the following steps: Step S1: Integrated Operating Condition Setup; Construct a general simulation test condition that simultaneously covers the variable speed fluctuation vibration of doubly-fed wind turbines (1000r / min-1800r / min) and the steady-state high-frequency vibration characteristics of direct-drive wind turbines (300r / min), achieving synchronous output of vibration simulation signals for both types of wind turbines. Connect the vibration acquisition unit under test to the standardized test link, complete the connection of the power supply circuit and signal transmission circuit, and perform link self-test to eliminate problems such as loose lines, signal interference, and link breaks, and lock the test hardware state to a constant state.
[0022] Step S2: Layered synchronous vibration sampling; Based on the differences in vibration characteristics between the doubly fed and direct-drive wind turbines, a differentiated synchronous sampling strategy is adopted. In step S2, the two sampling modes share the same signal acquisition channel and time axis, without the need to switch hardware links, ensuring that the data of the two types of wind turbines can be compared horizontally. Multi-dimensional raw vibration signals of the two types of wind turbines are collected in parallel to construct the raw vibration dataset.
[0023] In step S2, the differentiated sampling strategy for hierarchical synchronous vibration sampling is as follows: a continuous steady-state sampling mode is adopted for the low-frequency fluctuation vibration signal of the doubly fed wind turbine, and continuous acquisition is carried out without interruption throughout the process.
[0024] An intermittent encrypted sampling mode is adopted for high-frequency transient micro-vibration signals of direct-drive wind turbines. The sampling points for high-frequency abrupt changes are encrypted, and the two types of sampling are aligned in timing, correspond in measurement points, and executed synchronously and in parallel. Overall, continuous synchronous sampling lasts for 30 minutes, summarizing vibration data from both types of turbines across all dimensions to construct a complete original vibration dataset. This addresses the shortcomings of traditional step-by-step sampling and single-mode sampling, which cannot simultaneously capture the differentiated vibration signals of the two turbine models. It avoids the problems of missing low-frequency signals and distortion of high-frequency transient signals, while eliminating the need to switch hardware links, ensuring the consistency of horizontal comparison of test data between the two models, and significantly improving the completeness of data acquisition.
[0025] Step S3: Model quantization deviation compensation; pre-store the doubly-fed wind turbine speed reference threshold. B 1. Constant speed reference threshold for direct-drive fans B2. And a fixed model compensation coefficient K adapted to the fluctuation level of the operating conditions. After sampling, the preset model deviation compensation mathematical model is called to perform quantitative correction on the original vibration dataset group by group. In step S3, the model deviation compensation only performs algorithm calculations on the subsequent dataset and does not modify the original acquisition logic and operating parameters of the unit under test. The data offset caused by the inherent operating condition differences between the two types of wind turbines is eliminated, and the corrected vibration dataset is obtained.
[0026] It replaces the traditional rough correction method based on manual experience, and realizes full quantification, standardization and reproducibility of data correction. It completely solves the problem of test data deviation caused by the difference in the operating conditions of the two models, greatly improves the test accuracy from the bottom up, and does not interfere with the original working performance of the unit under test, thus ensuring the authenticity of the test.
[0027] In step S3, the mathematical model calculation formula for aircraft deviation compensation is as follows: .
[0028] in, X This represents the original vibration amplitude. Y The corrected vibration amplitude; K This refers to the model compensation coefficient; This is the baseline threshold for the model. Corresponding to doubly fed wind turbines; Corresponding to direct drive fans.
[0029] Step S4: Multi-level progressive performance discrimination; A two-level progressive verification logic is used to perform performance discrimination on the calibrated vibration dataset. In step S4, the two-level progressive performance discrimination logic includes: the first level is to verify the continuity, integrity and absence of missing parts of the vibration signal; the second level is to compare the consistency between the calibrated data and the standard sample data, and simultaneously verify the signal response delay index.
[0030] Based on the calibrated vibration dataset, signal integrity verification, acquisition accuracy and response performance determination are completed sequentially, and preliminary test results are output.
[0031] Layered screening of data defects involves filtering out invalid data first and then evaluating core performance. This avoids invalid data interfering with performance assessment results, solves the problems of low accuracy and missed defects in traditional single-shot general judgments, and improves the effectiveness and accuracy of preliminary judgment results.
[0032] Step S5: Constant Operating Condition Closed-Loop Retest; With the test operating parameters, sampling parameters, and environmental parameters kept constant, multiple retests are conducted on the initially abnormal samples. In Step S5, the closed-loop retest mechanism is as follows: For the initially marked abnormal samples, all test parameters are kept constant, and three consecutive closed-loop retests are performed. If only the three results show a consistent deviation, it is determined to be a true performance anomaly. Distinguish between random interference-induced pseudo-anomalies and true performance defects. If only a single deviation occurs while the other two results are normal, it is directly determined to be pseudo-anomaly data caused by random environmental interference and is discarded.
[0033] It effectively distinguishes between random interference and false anomalies and real equipment performance defects, solving the industry pain point of high misjudgment rate in traditional single-test discrimination, and significantly improving the reliability of test results. At the same time, it eliminates the need for repeated start-ups and shutdowns of equipment and switching of operating conditions, reducing hardware impact and wear.
[0034] Step S6: Environmental adaptive threshold correction; real-time environmental parameters are collected, and the real-time temperature of the test environment is collected throughout the test. T Spatial electromagnetic interference intensity E retrieve industry benchmark temperature T 0. Fixed temperature correction weight a 1. Electromagnetic interference correction weight a 2. The judgment threshold is dynamically updated through an environmental correction model, based on the threshold correction coefficient. The vibration performance judgment threshold is dynamically scaled and updated to offset the judgment error caused by environmental temperature and humidity fluctuations and electromagnetic interference, and the final standardized test result is output to complete the closed-loop test.
[0035] It enables dynamic adaptation of environmental parameters and judgment thresholds, solves the defect that fixed thresholds cannot adapt to environmental fluctuations, further eliminates the impact of external environmental interference on test results, and achieves high-precision standardized testing across all scenarios.
[0036] In step S6, the environmental adaptive correction is implemented based on the environmental correction model, the model formula of which is shown below: .
[0037] in, This is the threshold correction coefficient; Real-time temperature; Reference temperature; Electromagnetic interference intensity; and The adjustment weights are fixed.
[0038] After environmental re-inspection is completed, all valid samples will be integrated, along with all normal valid test samples and the confirmed abnormal samples. These samples will be categorized and labeled according to doubly-fed induction generator (DFIG) and direct-drive fan models, generating a standardized test report with dual-model classification labels, thus forming a complete testing loop. This achieves traceability of test data, standardized archiving of results, and adapts to the standardized requirements of industrial batch testing and factory calibration, thereby perfecting the overall testing closed-loop system.
[0039] Comparative Example 1 Using industry-standard testing methods, the test conditions for doubly-fed and direct-drive wind turbines were set up step by step, and signal acquisition was completed in two stages. Without a quantization compensation algorithm, data deviations were corrected by manual experience. A single data discrimination method was adopted, without closed-loop re-inspection and environmental adaptive correction processes, to complete the vibration acquisition performance test and record various test indicators.
[0040] Comparative Example 2 It adopts existing general automated testing methods and has basic data correction functions, but it is only configured with a single fixed threshold and lacks a dual-model differentiated compensation model; it lacks a hierarchical synchronous sampling mechanism and collects data step by step; it lacks environmental parameter linkage correction and multiple re-examination mechanisms to complete the vibration acquisition test of dual models and record the corresponding test data.
[0041] Four core evaluation indicators were standardized: data deviation rate (accuracy), single test duration, batch misjudgment rate, and equipment start-up and shutdown loss rate. All data were obtained from actual measurements under identical test conditions, and the data sources are authentic and reproducible. Specific comparison data are shown in the table below: Table 1. Quantitative Evaluation Indicators and Comparative Data
[0042] The data deviation rate is the percentage of the average deviation between the test data and the true value of the industry standard; the test duration is the average time spent on a complete test of two models; the batch misjudgment rate is the percentage of incorrect judgments among 100 batch test samples; the equipment start-up and shutdown loss rate is calculated based on 100% of the traditional manual test loss, and is converted according to the number of start-ups and shutdowns and hardware impact losses. All data are the average values of multiple actual measurements under the same working conditions, and are objective and repeatable.
[0043] Therefore, the present invention adopts the above-mentioned universal vibration monitoring and acquisition performance test method for doubly fed and direct-drive wind turbines. Through the optimization of full-quantization algorithm and closed-loop process, it effectively solves the shortcomings of traditional test technology, improves test accuracy, shortens the overall test time for dual-model turbines, reduces the misjudgment rate of batch tests, and reduces hardware losses during equipment start-up and shutdown.
[0044] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention; and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A universal vibration monitoring and acquisition performance testing method for both doubly-fed and direct-drive wind turbines, characterized in that, Includes the following steps: Step S1: Integrated working condition setup; Construct a general simulation test condition that simultaneously covers the variable speed fluctuation vibration of doubly fed wind turbines and the steady-state high-frequency vibration characteristics of direct-drive wind turbines, and connect the vibration acquisition unit under test to the standardized test link; Step S2: Layered synchronous vibration sampling; Based on the vibration differences between the two types of wind turbines, a differentiated synchronous sampling strategy was adopted to collect multi-dimensional original vibration signals of the two types of wind turbines in parallel and construct an original vibration dataset. Step S3: Model Quantization Deviation Compensation; Call the preset model deviation compensation mathematical model to quantize and correct the original vibration dataset, eliminate the data offset caused by the inherent operating condition differences between the two types of wind turbines, and obtain the corrected vibration dataset. Step S4: Multi-level progressive performance judgment; A two-level progressive verification logic is adopted. Based on the calibrated vibration dataset, the signal integrity verification, acquisition accuracy and response performance judgment are completed in sequence, and the preliminary test results are output. Step S5: Closed-loop re-inspection under constant operating conditions; By keeping the test conditions and environmental parameters unchanged, multiple retests were conducted on the initial abnormal samples to distinguish between random interference pseudo-anomalies and real performance defects. Step S6: Environmental Adaptive Threshold Correction; Collect real-time environmental parameters, dynamically update the judgment threshold through the environmental correction model, output the final standardized test result, and complete the closed-loop test.
2. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 1, characterized in that, In step S2, the differentiated sampling strategy for hierarchical synchronous vibration sampling is as follows: a continuous steady-state sampling mode is used for the low-frequency fluctuation vibration signal of the doubly fed fan, and an intermittent encrypted sampling mode is used for the high-frequency transient micro-vibration signal of the direct-drive fan. The two types of sampling are aligned in time, correspond in measurement points, and are executed synchronously in parallel.
3. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 2, characterized in that, In step S3, the mathematical model calculation formula for aircraft deviation compensation is as follows: ； in, X This represents the original vibration amplitude. Y The corrected vibration amplitude; K This refers to the model compensation coefficient; This is the baseline threshold for the model. Corresponding to doubly fed wind turbines; Corresponding to direct drive fans.
4. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 3, characterized in that, In step S4, the two-level progressive performance discrimination logic includes: the first level verifies the continuity, integrity and absence of missing parts of the vibration signal; the second level compares the consistency between the correction data and the standard sample data, and simultaneously verifies the signal response delay index.
5. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 4, characterized in that, In step S5, the closed-loop retesting mechanism is as follows: for the initially marked abnormal samples, all test parameters are kept constant, and three consecutive closed-loop retests are performed. If the results of the three tests are consistent and deviate, it is determined to be a true performance abnormality.
6. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 5, characterized in that, In step S6, the environmental adaptive correction is implemented based on the environmental correction model, the model formula of which is shown below: ； in, This is the threshold correction coefficient; Real-time temperature; Reference temperature; Electromagnetic interference intensity; and The adjustment weights are fixed.
7. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 6, characterized in that, In step S2, the two sampling modes share the same signal acquisition channel and time axis, without the need to switch hardware links, ensuring that the data of the two models can be compared horizontally.
8. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 7, characterized in that, In step S3, the model deviation compensation only performs algorithm calculations on the post-dataset and does not modify the original acquisition logic and operating parameters of the unit under test.
9. The method for testing the vibration monitoring and acquisition performance of both doubly-fed and direct-drive wind turbines according to claim 8, characterized in that, In step S5, after the re-inspection is completed, all valid samples are integrated to generate a standardized test report with dual model classification labels, forming a complete test closed loop.