Offshore wind turbine generator health degree evaluation method based on multi-source data fusion

Abstract

The invention belongs to the technical field of offshore wind turbine monitoring, and particularly provides an offshore wind turbine generator health degree evaluation method based on multi-source data fusion, and the method comprises the following steps: 1, constructing a health scoring model; 2, performing health degree scoring on each module of the offshore wind turbine generator; the health degree scores, calculated in the second step, of all modules of the offshore wind turbine generator set are input into the health scoring model in the first step, a total score is obtained, and then health state grading is carried out. According to the method, multiple indexes of the fan are comprehensively considered, and the health state of the unit is evaluated in real time.

Description

Technical Field

[0001] The present invention belongs to the technical field of offshore wind turbine monitoring. Specifically, it relates to a method for evaluating the health of offshore wind turbines based on multi-source data fusion. Background Art

[0002] Offshore wind turbines operate in an environment of high salt, high humidity, and strong winds. The health of the wind turbines directly affects their power generation efficiency and safety. Wind turbines not only have traditional mechanical, electrical, and structural components, but also involve marine engineering problems (such as foundation scouring, submarine cable wear, etc.) and the quality of lubricating oil and insulating oil. These factors will gradually accumulate with the long-term operation of the unit, affecting the performance and service life of the wind turbine.

[0003] However, the current methods for evaluating the health of offshore wind turbines have the following deficiencies: 1. Data island phenomenon: Data between different monitoring systems (such as SCADA system, CMS system, oil quality monitoring, video monitoring, etc.) cannot be effectively fused, and the health status of the wind turbine cannot be comprehensively evaluated.

[0004] 2. Ignoring the influence of environmental factors: The influence of environmental changes (such as strong winds, submarine cable wear, etc.) on the wind turbine has not been fully considered.

[0005] 3. The influence of over-generation on the life of the wind turbine is not included in the evaluation: Long-term overload operation will accelerate the wear of wind turbine components and affect the life of the wind turbine.

[0006] 4. Lack of a dynamic health assessment mechanism: Existing assessment methods cannot adjust the assessment model according to real-time monitoring data, resulting in a lag in early warning.

[0007] To make up for these deficiencies, the present invention proposes a health evaluation method based on multi-source data fusion, comprehensively considering multiple indicators of the wind turbine, such as mechanical, electrical, structural, fire protection, work order issuance for defect elimination (reverse indicator), environment, oil products, and operation efficiency, to real-time evaluate the health status of the unit, timely discover potential problems, and optimize the operation and maintenance management of the wind turbine. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide a method for evaluating the health of offshore wind turbines based on multi-source data fusion, comprehensively considering multiple indicators of the wind turbine, and real-time evaluating the health status of the unit.

[0009] To solve the above technical problem, the technical solution adopted by the present invention is: A method for evaluating the health of offshore wind turbines based on multi-source data fusion, comprising the following steps: Step 1. Construct a health scoring model: By integrating multi-source data of offshore wind turbines, a comprehensive health scoring model is formed. The calculation formula of the health scoring model is as follows: ; In the formula: H mech represents the health score of the mechanical part; H elec represents the health score of the electrical part; H struc represents the health score of the structural part; H env represents the health score of the environmental part; H oil represents the health score of lubricating oil and insulating oil; H ops represents the operating efficiency score; H video represents the status score of the video monitoring system; H overload represents the impact score of over - generation on the life of the fan; H temp represents the health score of the internal temperature of the fan; H fire represents the health score of the fire protection system; H ticket represents the defect elimination work order frequency score; ω * represents the weight coefficient of each module, which is dynamically adjusted according to the contribution of different modules to the health of the fan; Step 2: Conduct a health assessment of each module of the offshore wind turbine; Step 3: Comprehensive scoring and grading: Input the health scores of each module of the offshore wind turbine calculated in Step 2 into the health scoring model in Step 1 to obtain the total score, and then conduct a health status level division.

[0010] In the preferred solution, the health score of the mechanical part H mech is obtained through the following steps: 1) Feature extraction Use the fast Fourier transform to extract the vibration spectrum of the mechanical components of the fan and identify early fault features. The expression is: ; Envelope analysis is used to extract the impact signal features of gears and bearings. The expression is: ; In the formula, X ( t ) represents the vibration spectrum of the mechanical component vibration signal of the fan, j represents the phase component of the vibration signal, f represents the frequency of the vibration signal; 2) Calculation of the health score of the mechanical part Extract key features from the spectrum and envelope signals, including the amplitude of specific fault frequencies, the peak or root mean square value of the envelope signal, and normalize the key features to obtain the health indicators of each component Ri The expression is: ; Calculate the health score of the mechanical part: ; Among them, R i is the health indicator of each mechanical component, R max is the maximum health value of each mechanical component.

[0011] In the preferred solution, the health score of the electrical part H elec is obtained through the following steps: 1) Current harmonic analysis Calculate the total harmonic distortion rate THD of the current to evaluate the health of the electrical system, THD The higher it is, the more serious the distortion of the current waveform. The calculation formula is: ; In the formula: I 1 represents the effective value or amplitude of the first harmonic current, I n represents the n th harmonic current effective value or amplitude; 2) Voltage unbalance Calculate the voltage unbalance to judge whether the electrical system is normal. The calculation formula is: ; In the formula, U unb represents the voltage unbalance, max(U a ,U b ,U c ) represents the three-phase voltage U a 、U b 、U c in the maximum value, min(U a ,U b ,U c ) represents the minimum value in the three-phase voltage, U avgRepresents the average value of the three-phase voltage; U avg The calculation formula is: ; 3) The electrical part health score will THD and U unb After normalizing by the maximum value and adding them together, the calculation formula is: ; In the formula, THD max , U unb,max Respectively represent the maximum allowable values of the total harmonic distortion rate and the voltage unbalance degree; In the preferred solution, the structural part health score H struc Is obtained through the following steps: 1) The expression of the influence of the change of the foundation scour depth over time on the stability of the fan is: ; In the formula, D scour Represents the actually measured foundation scour depth; D max Represents the maximum allowable scour depth of the design; 2) The expression of the influence of strong wind weather on the foundation stability is: ; In the formula, W wind Represents the actual wind speed, W max Represents the maximum wind resistance of the fan structure design; 3) The expression of evaluating the influence of the aging of the anti-corrosion layer on the reliability of the fan is: ; In the formula, R c Represents the current aging degree of the anti-corrosion layer, R max Represents the maximum allowable aging degree of the anti-corrosion layer; 4) The expression of the influence of submarine cable wear on the stability of the fan is: ; In the formula, M wear Represents the actual wear amount of the submarine cable, M max Represents the maximum allowable wear amount of the submarine cable; 5) Comprehensive structure score, the expression is: H struc =γ·H scour +δ·H stability +ε·H corrosion +ζ·H cable ; Wherein, γ, δ, ε, ζ are The weight coefficients respectively represent the contribution ratios of foundation scouring, strong wind influence, anti-corrosion aging, and submarine cable abrasion to the total score, γ + δ + ε + ζ = 1.

[0012] In the preferred solution, the health score of the video monitoring H video is obtained through the following steps: 1) Calculation of the camera availability: ; 2) Evaluation of the image clarity: The Laplacian variance method is used to evaluate the I image clarity, and the expression is: ; Wherein, represents the Laplacian operator; 3) Comprehensive health score of the video monitoring: ; Wherein: α and β are weight coefficients set according to actual requirements, α + β = 1, n represents the number of images participating in the calculation.

[0013] In the preferred solution, the operation efficiency score H ops is obtained through the following steps: 1) Availability U avail Calculation: ; 2) Equivalent utilization hours U eff Calculation: ; 3) Operation efficiency score H ops Calculation: ; In the formula, ω 1 and ω 2 are weights determined according to the actual situation, and ω 1 +ω 2 = 1, U eff，max represents the set reference value of the maximum equivalent utilization hours.

[0014] In a preferred embodiment, the oil quality detection and health score H oil is obtained through the following steps: 1) Lubricating oil detection Lubricating oil quality H oil,lube evaluation: ; In the formula: P i represents the actual measured value of each detection index of the lubricating oil, P max represents the maximum allowable value of each corresponding index of the lubricating oil; 2) Box-type transformer insulating oil detection Evaluate the state of the box-type transformer insulating oil through the breakdown voltage of the insulating oil H oil,ins : ; In the formula: Δ V represents the decrease value of the breakdown voltage of the insulating oil, that is, the decrease amplitude of the current breakdown voltage compared with the initial or reference value, Vmax represents the maximum reference value of the breakdown voltage of the insulating oil; 3) Comprehensive oil quality health score ; In the formula: α and β are weight coefficients, respectively representing the importance of the lubricating oil quality score and the box-type transformer insulating oil state score in the comprehensive oil quality health score, and α + β = 1 。

[0015] In a preferred embodiment, the influence score of the fan over-generation H overload is calculated by the following formula: ; In the formula: P overload represents the actual power when the fan operates overloaded, P rated represents the rated power of the fan; In the preferred solution, the temperature monitoring and health score H temp is calculated by the following formula: ; wherein, T current is the current unit temperature, T max is the maximum allowable temperature.

[0016] In the preferred solution, the health score of the fire protection system H fire is obtained by evaluating the monitoring data of the fire protection system to determine whether the fire protection system is in normal working condition. The calculation formula is: ; wherein, F fail is the number of failures of the fire protection system, F max is the maximum allowable number of failures.

[0017] In the preferred solution, the impact score of the defect elimination work ticket H ticket The calculation formula is: ; wherein, N ticket is the number of defect elimination work tickets issued per unit time, N max is the maximum allowable number, k penalty is the penalty factor affected by the work ticket.

[0018] In the preferred solution, the health score of the environmental part H env The calculation formula is: ; wherein: F wind , F temp , F humidity , F corrosion , F lightning are environmental impact factors, which are wind speed impact factor, temperature impact factor, humidity impact factor, salt spray corrosion impact factor and lightning strike frequency impact factor respectively. α1, α2, α3, α4, α5 are the weights of each environmental impact factor.

[0019] In a preferred embodiment, in step three, the health status is classified according to the following levels: Level A, healthy: H ≥ 90; Level B, good: 75 ≤ H <90; Level C, fair: 60 ≤ H <75; Level D, warning: 40 ≤ H <60; Level E, dangerous: H <40.

[0020] A method for evaluating the health of an offshore wind turbine based on multi-source data fusion provided by the present invention has the following beneficial effects: 1. Comprehensive health assessment: covering multiple aspects of indicators such as machinery, electricity, structure, fire protection system, environment, oil products, over-generation, and defect elimination work tickets, providing a comprehensive health assessment.

[0021] 2. Dynamic adjustment mechanism: adjusts the weights and scores of each module according to real-time data, improving the real-time performance and accuracy of the assessment.

[0022] 3. Refined health scoring: quantifies the scoring and hierarchical management according to the influence of different factors on the health of the fan.

[0023] 4. Precise identification of early problems: discovers potential faults of the fan in advance through data fusion and refined analysis.

[0024] 5. Efficient health scoring system: adopts a percentile scoring system, which is convenient for real-time monitoring, hierarchical management, and decision support. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The present invention will be further described below with reference to the drawings and embodiments: Figure 1 It is a schematic flow chart of the implementation steps of the method of the present invention; Figure 2 It is a radar chart of the module scores of the unit in Embodiment 2; Figure 3 It is a monthly trend chart of the unit health score in Embodiment 2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0027] Embodiment 1: As Figure 1As shown in the figure, a method for evaluating the health of an offshore wind turbine based on multi-source data fusion is characterized by the following steps: Step 1. Construct a health scoring model: By integrating multi-source data of the offshore wind turbine, a comprehensive health scoring model is formed. The calculation formula of the health scoring model is as follows: ; In the formula: H mech represents the health score of the mechanical part; H elec represents the health score of the electrical part; H struc represents the health score of the structural part; H env represents the health score of the environmental part; H oil represents the health score of lubricating oil and insulating oil; H ops represents the operation efficiency score; H video represents the status score of the video monitoring system; H overload represents the score of the impact of over-generation on the life of the fan; H temp represents the health score of the internal temperature of the fan; H fire represents the health score of the fire protection system; H ticket represents the score of the frequency of defect elimination work orders; ω * represents the weight coefficient of each module, which is dynamically adjusted according to the contribution of different modules to the health of the fan.

[0028] Specifically, in this embodiment, the weight coefficients of each module are given an initial weight allocation according to industry standards, expert experience, and past research data. In different stages of the fan life cycle (such as early operation, mid-term maintenance, aging stage), the degree of influence of each module on health may change. Therefore, the weights can be set to be dynamically adjusted according to time or the state of the fan.

[0029] Step 2. Score the health of each module of the offshore wind turbine.

[0030] 1. Health score of the mechanical part H mech Obtained through the following steps: 1) Feature extraction Use the fast Fourier transform to extract the vibration spectrum of the fan's mechanical components and identify early fault features. The expression is: ; Envelope analysis is used to extract the impact signal characteristics of gears and bearings, and the expression is: ; In the formula, X ( t ) represents the frequency spectrum of the vibration signal of the mechanical components of the fan, j represents the phase component of the vibration signal, f represents the frequency of the vibration signal; 2) Calculation of the health score of the mechanical part Extract key features from the frequency spectrum and envelope signal, including the amplitude of specific fault frequencies, the peak or root mean square value of the envelope signal, and normalize the key features to obtain the health indicators of each component Ri The expression is: ; Calculate the health score of the mechanical part: ; Among them, R i is the health indicator of each mechanical component, R max is the historical maximum fault value or maximum allowable value of each mechanical component.

[0031] 2. Health score of the electrical part H elec Obtained through the following steps: 1) Current harmonic analysis Calculate the total harmonic distortion rate THD of the current to evaluate the health of the electrical system. THD The higher it is, the more serious the distortion of the current waveform. The calculation formula is: ; In the formula: I 1 represents the effective value or amplitude of the current of the first harmonic, I n represents the n th harmonic's effective value or amplitude of the current.

[0032] Excessive harmonic content will cause phenomena such as equipment overheating, power quality degradation, affecting the stability of the fan control system, and increasing the losses of electrical equipment.

[0033] 2) Voltage unbalance: Ideally, the three-phase voltages should be equal. However, due to the power fluctuations of the wind turbine, load imbalance, or grid problems, voltage unbalance may occur.

[0034] Calculate the voltage unbalance to judge whether the electrical system is normal. The calculation formula is: ; In the formula,U unb Indicates the voltage unbalance degree max(U a ,U b ,U c ) Indicates the three-phase voltage U a 、U b 、U c The maximum value in min(U a ,U b ,U c ) Indicates the minimum value in the three-phase voltage U avg Indicates the average value of the three-phase voltage; U avg The calculation formula of is: ; 3) The electrical part health score will THD And U unb After being normalized by the maximum value and added together, the calculation formula is: ; In the formula, THD max , U unb,max Respectively represent the maximum allowable values of the total harmonic distortion rate and the voltage unbalance degree; IEEE 519-2014 standard: The THD of the low-voltage side of the wind turbine generator set should be ≤ 5%; IEC 60034-1 (international standard): The general requirement for the voltage unbalance degree of the motor is ≤ 2%.

[0035] Therefore, in this embodiment, THD max = 5%, U unb,max = 2%.

[0036] If THD = 0 and Uunb = 0, then Helec = 100%, indicating that the electrical system is completely healthy.

[0037] If THD = THD max And U unb = U ​unb,max , then Helec = 0%, indicating a serious system failure.

[0038] 3. Structural partial health score H struc Obtained through the following steps: 1) The expression for the impact of the change in the foundation scour depth over time on the stability of the wind turbine is: ; In the formula, D scour represents the actually measured foundation scour depth; D max represents the maximum allowable scour depth in the design; 2) The expression for the impact of strong wind weather on the foundation stability is: ; In the formula, W wind represents the actual wind speed, W max represents the maximum wind resistance of the wind turbine structure design, that is, the safety threshold under extreme conditions. Usually provided by the wind farm planning and design institute or the wind turbine manufacturer, considering the wind resistance strength of the foundation structure.

[0039] When W wind / W max is close to 1, it indicates that the wind turbine is facing a large wind pressure impact and the foundation safety is reduced.

[0040] 3) The expression for evaluating the impact of the anti-corrosion layer aging on the reliability of the wind turbine is: ; In the formula, R c represents the current aging degree of the anti-corrosion layer, R max represents the maximum allowable aging degree of the anti-corrosion layer.

[0041] The aging of the anti-corrosion layer will cause the corrosion of the steel structure, thereby reducing the overall reliability of the wind turbine. The smaller this ratio, the higher the corrosion risk and the lower the health score.

[0042] 4) The expression for the impact of submarine cable abrasion on the stability of the wind turbine is: ; In the formula, M wear represents the actual abrasion amount of the submarine cable, which can be the relative value of the abrasion depth, damage index or remaining life. Usually obtained through ROV (Remotely Operated Vehicle) or cable condition monitoring system.

[0043] M max It represents the maximum allowable wear of the submarine cable. Exceeding this value may lead to accidents such as insulation failure, leakage, and cable breakage.

[0044] The long-term pulling effect of cable moisture, wind and waves, catenary, etc. will accelerate the wear, affecting the power transmission reliability and structural stability of the whole machine.

[0045] 5) Comprehensive structure score, the expression is: H struc =γ·H scour +δ·H stability +ε·H corrosion +ζ·H cable ; In the formula, γ, δ, ε, ζ are Weight coefficients, respectively representing the contribution ratios of foundation scour, strong wind influence, anti-corrosion aging, and submarine cable wear to the total score, γ + δ + ε + ζ = 1.

[0046] In this embodiment, the values of each weight coefficient are referred to in Table 1.

[0047]

[0048] 4. Video monitoring health score H video Obtained through the following steps: 1) Camera availability calculation: ; 2) Image clarity evaluation: Use the Laplacian variance method to evaluate the I image clarity, the expression is: ; In the formula, represents the Laplacian operator; 3) Video monitoring health comprehensive score: ; In the formula: α and β are weight coefficients set according to actual needs, α + β = 1, n represents the number of images participating in the calculation.

[0049] In this embodiment, α The value range is 0.5 to 0.7, β The value range is 0.3 to 0.5.

[0050] 5. Operating efficiency score H ops Obtained through the following steps: 1) Availability U avail Calculate: ; 2) Equivalent utilization hours U eff Calculate: ; 3) Operating efficiency score H ops Calculate: ; In the formula, ω 1 and ω 2 are weights determined according to the actual situation, and ω 1 + ω 2 = 1. U eff，max Represents the set reference value of the maximum equivalent utilization hours, such as the average value, theoretical value or historical maximum value of the wind turbines with the best wind resources in this area.

[0051] ω 1 and ω 2 The value range is 0.4 to 0.6. In areas with large wind resource differences, ω 2 weight can be appropriately increased to highlight the efficiency difference.

[0052] 6. Oil product detection and health score H oil Obtained through the following steps: 1) Lubricating oil detection Lubricating oil quality H oil,lube Evaluate: ; In the formula: P i Represents the actual measured values of various detection indicators of the lubricating oil, such as acid value, viscosity, moisture, particle contamination degree, metal wear particle concentration, etc.

[0053] P max Represents the maximum allowable value of the corresponding indicators of the lubricating oil. Exceeding this value is regarded as the degradation or failure of the lubricating oil performance, as shown in Table 2 for reference.

[0054] WhenP i is small, indicating that the oil product is in good condition H oil,lube approaches 1

[0055] When P i → P max , it indicates that this index is approaching the failure boundary and the score decreases

[0056] If P i > P max , the score is usually set to 0 or an abnormal alarm is triggered

[0057] 2) Detection of insulating oil in box-type substations Evaluate the condition of the insulating oil in the box-type substation through the breakdown voltage of the insulating oil H oil,ins : ; In the formula: Δ V represents the decrease value of the breakdown voltage of the insulating oil, that is, the decrease amplitude of the current breakdown voltage compared with the initial or reference value Vmax represents the maximum reference value of the breakdown voltage of the insulating oil

[0058] The maximum reference value of the breakdown voltage (which can be regarded as the rated breakdown voltage in the initial healthy state), for example: 60 kV, 70 kV, etc. It is determined by the equipment manufacturer or the detection data in the initial operation period and used as the reference upper limit for normalization

[0059] 3) Comprehensive oil product health score ; In the formula: α and β are weight coefficients, representing the importance of the lubricating oil quality score and the box-type substation insulating oil condition score in the comprehensive oil product health score respectively, and α + β = 1 。

[0060] α and β The recommended ratio of is shown in Table 2

[0061]

[0062] 7. Influence score of over-generation of the fan H overload Calculated by the following formula: ; In the formula: Poverload Indicates the actual power when the fan is operating at overload. P rated Indicates the rated power of the fan.

[0063] 8. Temperature Monitoring and Health Scoring H temp Calculated by the following formula: ; Where T current is the current unit temperature, T max is the maximum allowable temperature.

[0064] 9. Fire Protection System Health Scoring H fire By evaluating the monitoring data of the fire protection system, determine whether the fire protection system is in normal working condition. The calculation formula is: ; Where F fail is the number of faults of the fire protection system, F max is the maximum allowable number of faults.

[0065] 10. Impact Scoring of Defect Elimination Work Tickets H ticket The calculation formula is: ; Where N ticket is the number of defect elimination work tickets issued per unit time, N max is the maximum allowable number, k penalty is the penalty factor affected by the work ticket.

[0066] 11. Environmental Part Health Scoring H env The calculation formula is: ; Where: F wind , F temp , F humidity , F corrosion , F lightningis the environmental impact factor, which are the wind speed impact factor, temperature impact factor, humidity impact factor, salt spray corrosion impact factor and lightning strike frequency impact factor respectively. α1, α2, α3, α4, α5 are the weights of each environmental impact factor.

[0067] 1) Wind speed impact factor F wind The calculation formula is: ; In the formula, V max is the design limit wind speed, V is the actual wind speed.

[0068] The closer the wind speed is to the limit value, the greater the pressure on the fan structure, and the lower the score.

[0069] 2) F temp The calculation formula of is as follows: ; In the formula, is the most suitable temperature for the working environment, T is the measured temperature, represents the temperature fluctuation. The smaller the fluctuation, the closer the temperature impact factor is to 100 points.

[0070] F humidity The calculation formula of is as follows: ; In the formula, is the critical humidity, is the measured humidity.

[0071] The larger, the smaller the environmental humidity, and the more conducive to the operation of the equipment.

[0072] F corrosion The calculation formula of is as follows: ; In the formula, represents the salt spray deposition rate threshold, represents the salt spray deposition rate.

[0073] F lightning The calculation formula of is as follows: ; In the formula, represents the maximum number of lightning strikes resistant within the statistical time, and N represents the number of lightning strikes within the statistical time.

[0074] An example is given as follows: Environmental data of offshore wind farm: Fwind = 30, Ftemp = 10, Fhumidity = 60, Fcorrosion = 80, Flightning = 40.

[0075] Weight α 1 = 0.2, α 2 = 0.1, α 3 = 0.1, α 4 = 0.4, α 5 = 0.2.

[0076] Calculation H en = 0.2×30 + 0.1×10 + 0.1×60 + 0.4×80 + 0.2×40 = 53 points.

[0077] α1, α2, α3, α4, α5 are dynamically adjusted through historical data, and the operation data and fault records of wind turbines under different environmental conditions are collected. For example, if it is found that the number of faults caused by lightning strikes is relatively large in history and has a greater impact on the health of the unit, then the weight of the lightning strike frequency impact factor α 5 can be appropriately increased; conversely, if the problems caused by a certain environmental factor are fewer, its weight can be reduced.

[0078] Step 3. Comprehensive scoring and grading: Input the health score of each module of the offshore wind turbine calculated in Step 2 into the health scoring model in Step 1 to obtain the total score, and then conduct a health status level division.

[0079] The health status is divided according to the following levels: Grade A, healthy: H ≥ 90; Grade B, good: 75 ≤ H < 90; Grade C, average: 60 ≤ H < 75; Grade D, warning: 40 ≤ H < 60; Grade E, dangerous: H < 40.

[0080] Example 2: To verify the effectiveness of the "offshore wind turbine health assessment method based on multi-source data fusion" described in the present invention, three offshore wind turbines (numbered WT-01, WT-02, WT-03) located in a wind farm on the east coast of China were selected as test objects, and their operation data were continuously monitored and analyzed for 6 months. The specific content is as follows: The radar chart of the unit module score is shown in Figure 2 , and the monthly trend chart of the unit health score is shown in Figure 3 .

[0081] In April 2024, the H mech of WT-02 dropped suddenly from 83 to 71. Through FFT and envelope analysis, it was found that the characteristic frequency of the inner ring of the bearing was abnormal. After replacing the faulty parts in advance, the risk of shutdown was avoided.

[0082] The comparison of operation and maintenance indicators before and after implementation is shown in Table 3.

[0083]

[0084] Those skilled in the art can easily understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for evaluating the health status of an offshore wind turbine based on multi-source data fusion, characterized in that, It includes the following steps: Step 1. Construct a health score model: By integrating multi-source data of an offshore wind turbine, a comprehensive health score model is formed. The calculation formula of the health score model is as follows: ； Wherein: H mech represents the health score of the mechanical part; H elec represents the health score of the electrical part; H struc represents the health score of the structural part; H env represents the health score of the environmental part; H oil represents the health score of lubricating oil and insulating oil; H ops represents the operating efficiency score; H video represents the status score of the video monitoring system; H overload represents the score of the impact of over - generation on the fan life; H temp represents the health score of the internal temperature of the fan; H fire represents the health score of the fire protection system; H ticket represents the frequency score of the defect elimination work ticket; ω * represents the weight coefficient of each module, which is dynamically adjusted according to the contribution of different modules to the fan health; Step 2. Score the health of each module of the offshore wind turbine; Step 3. Comprehensive scoring and grading: Input the health scores of each module of the offshore wind turbine calculated in Step 2 into the health score model in Step 1 to obtain the total score, and then conduct health status level division.

2. The health assessment method of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The mechanical part health score H mech is obtained through the following steps: 1) Feature extraction Use the fast Fourier transform to extract the vibration spectrum of the mechanical components of the fan, and identify early fault features. The expression is: ； Envelope analysis is used to extract the impact signal features of gears and bearings. The expression is: ； In the formula, X ( t ) represents the frequency spectrum of the vibration signal of the fan mechanical components, j represents the phase component of the vibration signal, f represents the frequency of the vibration signal; 2) Calculation of the health score of the mechanical part Extract key features from the frequency spectrum and envelope signals, including the amplitudes of specific fault frequencies, the peak values or root mean square values of envelope signals, normalize the key features, and obtain the health indicators of each component Ri The expression is as follows: ； Calculate the health score of the mechanical part: ； Among them, R i are the health indicators of each mechanical component, R max are the maximum health values of each mechanical component.

3. The health assessment method of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The health score of the electrical part H elec is obtained through the following steps: 1) Current harmonic analysis Calculate the total harmonic distortion (THD) of the current to evaluate the health of the electrical system. THD The higher it is, the more serious the distortion of the current waveform. The calculation formula is as follows: ； In the formula: I 1 represents the effective value or amplitude of the current of the first harmonic, I n represents the n effective value or amplitude of the current of the th harmonic; 2) Voltage unbalance degree Calculate the voltage unbalance degree to judge whether the electrical system is normal. The calculation formula is: ； In the formula, U unb represents the voltage unbalance degree, max(U a ,U b ,U c ) represents the maximum value of the three-phase voltage U a 、U b 、U c in it, min(U a , U b ,U c ) represents the minimum value of the three-phase voltage, U avg represents the average value of the three-phase voltage; U avg The calculation formula is as follows: ; 3) The health score of the electrical part will be calculated by adding the normalized values of THD THD and U unb after normalizing them by their maximum values. The calculation formula is as follows: ； In the formula, THD max and U unb,max respectively represent the maximum allowable values of the total harmonic distortion rate and the voltage unbalance degree.

4. A method for evaluating the health of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that, The health score of the described structural part H struc is obtained through the following steps: 1) The expression for the influence of the change of the foundation scouring depth over time on the stability of the fan is: ； In the formula, D scour represents the actually measured foundation scour depth; D max represents the maximum allowable scour depth in design; 2) The expression for the influence of strong wind weather on the foundation stability is: ； In the formula, W wind represents the actual wind speed, W max represents the maximum wind resistance of the wind turbine structure design; 3) The expression for evaluating the influence of the aging of the anti-corrosion layer on the reliability of the fan is: ； In the formula, R c represents the current aging degree of the anti-corrosion layer, R max represents the maximum allowable aging degree of the anti-corrosion layer; 4) The expression for the influence of submarine cable abrasion on the stability of the fan is: ； In the formula, M wear represents the actual wear amount of the submarine cable, M max represents the maximum allowable wear amount of the submarine cable; 5) Comprehensive structure scoring. The expression is: H struc =γ·H scour +δ·H stability +ε·H corrosion +ζ·H cable ； In the formula, γ, δ, ε, ζ are Weight coefficients, respectively representing the contribution ratios of foundation scouring, strong wind influence, anti-corrosion aging, and submarine cable abrasion to the total score, γ + δ + ε + ζ = 1.

5. The method for evaluating the health degree of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The video surveillance health score H video is obtained through the following steps: 1) Calculation of the availability rate of the camera: ； 2) Evaluation of image clarity: Evaluating the clarity of an image using the Laplacian variance method, with the expression: I ​ ； In the formula, represents the Laplace operator; 3) Comprehensive health scoring of video monitoring: ； Wherein: α and β are weight coefficients set according to actual requirements, α + β = 1, n represents the number of images participating in the calculation.

6. The method for evaluating the health of an offshore wind turbine based on multi-source data fusion according to claim 1, wherein The operating efficiency score H ops is obtained through the following steps: 1) Availability U avail Calculation: ； 2) Equivalent utilization hours U eff Calculation: ； 3) Operating efficiency score H ops Calculate: ； In the formula, ω 1 and ω 2 are weights determined according to the actual situation, and ω 1 + ω 2 = 1, U eff，max represents the reference value of the set maximum equivalent utilization hours.

7. A method for evaluating the health of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The oil product detection and health score H oil Obtained through the following steps: 1) Lubricating oil detection Lubricating oil quality H oil,lube Evaluation: ； Wherein: P i represents the actual measured value of each detection index of the lubricating oil; P max represents the maximum allowable value of the lubricating oil corresponding to each index; 2) Insulating oil detection of the box-type substation Evaluating the Insulating Oil Condition of Box-Type Substations through the Breakdown Voltage of Insulating Oil H oil,ins : ； where: Δ V represents the decrease value of the breakdown voltage of the insulating oil, that is, the decrease amplitude of the current breakdown voltage compared with the initial or reference value, Vmax represents the maximum reference value of the breakdown voltage of the insulating oil; 3) Comprehensive health scoring of the oil products ； Wherein: α and β are weight coefficients, respectively representing the importance of the lubricating oil quality score and the box-type substation insulating oil state score in the comprehensive oil product health score, and α + β = 1 。。 8. A method for evaluating the health status of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The influence score of the over-generation of the fan H overload Calculated by the following formula: ； In the formula: P overload represents the actual power when the fan operates under overload, P rated represents the rated power of the fan.

9. The method for evaluating the health degree of an offshore wind turbine based on multi-source data fusion according to claim 1, wherein, The temperature monitoring and health score H temp is calculated by the following formula: ； Among them, T current is the current unit temperature, T max is the maximum allowable temperature.

10. A method for evaluating the health status of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that The health score of the fire protection system H fire By evaluating the monitoring data of the fire protection system, it is determined whether the fire protection system is in a normal working state. The calculation formula is as follows: ； Among them, F fail is the number of failures of the fire protection system, F max is the maximum allowable number of failures.

11. A method for evaluating the health of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that, The defect elimination work order affects the scoring H ticket The calculation formula is as follows: ； Among them, N ticket is the number of defect elimination work tickets issued per unit time, N max is the maximum allowed number, k penalty is the penalty factor affected by the work ticket.

12. A method for evaluating the health of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that, The environmental partial health score H env The calculation formula is as follows: ； Wherein: F wind , F temp , F humidity , F corrosion , F lightning are environmental impact factors, namely the wind speed impact factor, the temperature impact factor, the humidity impact factor, the salt spray corrosion impact factor and the lightning strike frequency impact factor respectively, and α1, α2, α3, α4, α5 are the weights of each environmental impact factor.

13. A method for evaluating the health degree of an offshore wind turbine based on multi-source data fusion according to claim 1, characterized in that, In the above Step 3, the health status is divided according to the following levels: Level A, Healthy: H ≥ 90; Level B, Good: 75 ≤ H < 90; Grade C, General: 60 ≤ H < 75; Class D, Warning: 40 ≤ H < 60; Level E, Danger: H < 40.

Images