A wind farm operation and maintenance system

Abstract

This invention belongs to the field of operation and maintenance system technology, specifically relating to a wind farm operation and maintenance system, including a data acquisition module, a data processing module, a fault analysis and tracing module, a new project construction assistance module, and a design and assembly scheme update prompt module. Through the synergistic effect of the fault analysis and tracing module, the new project construction assistance module, and the design and assembly scheme update prompt module, this system significantly improves the operation and maintenance efficiency, design reliability, and equipment lifespan of wind farms. This system, through a data-driven closed-loop management mechanism, deeply integrates fault analysis, design optimization, and dynamic updates, overcoming the limitations of traditional wind farm operation and maintenance systems that rely on manual experience.

Description

Technical Field

[0001] This invention belongs to the field of operation and maintenance system technology, and specifically relates to a wind farm operation and maintenance system. Background Technology

[0002] The operation and maintenance of existing wind farms face numerous challenges, mainly including the following aspects:

[0003] 1. Wind farm equipment generates a large amount of operational data, but existing systems lack the ability to analyze and utilize this data, failing to fully realize its value in order to optimize equipment operation and maintenance strategies.

[0004] 2. New wind farm projects often face uncertainties in the design and construction phases, lack analysis and utilization of historical failure data, and are prone to repeating past mistakes, leading to similar problems in new projects.

[0005] 3. Even with new designs, assembly schemes, or new materials, existing systems lack effective mechanisms to update and apply these improvements in a timely manner, preventing equipment from benefiting from the latest technological advancements.

[0006] Existing technologies mainly monitor equipment status through sensors, use acoustic data for fault early warning, and perform fault prediction based on multi-source data. However, most of these systems focus on the implementation of a single function and lack comprehensive management capabilities for the entire process of wind farm operation and maintenance, and cannot solve the aforementioned technical problems.

[0007] For example, the invention patent with patent publication number CN113836762B discloses a wind turbine and wind farm digital mirror simulation display system, which focuses on realizing the comprehensive display and optimized control of wind farms through digital twin technology.

[0008] The invention patent with patent publication number CN114611424B discloses a method for predicting the life of large wind turbine blades by integrating acoustic data and CAE algorithms. Its core lies in life prediction and simulation analysis, but it also cannot solve the above-mentioned technical problems. Summary of the Invention

[0009] To address the aforementioned technical problems, this invention provides a wind farm operation and maintenance system. This system significantly improves the operation and maintenance efficiency, design reliability, and equipment lifespan of wind farms through the synergistic effect of a fault analysis and tracing module, a new project construction assistance module, and a design and assembly scheme update prompt module.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0011] A wind farm operation and maintenance system, comprising:

[0012] Data acquisition module: includes wind speed and direction sensors, wind turbine blade stress sensing devices, vibration sensors, microphone arrays and salt spray concentration sensors, used to collect wind conditions, equipment stress, vibration, acoustic data and environmental salt spray concentration in real time;

[0013] Data processing module: After preprocessing the collected data, it generates a device health score through a multi-parameter fusion algorithm;

[0014] Fault Analysis and Traceability Module: Based on the random forest model, it distinguishes between internal and external factors of failure, constructs a cause-effect graph for internal factors to trace the root causes of design defects, manufacturing errors or component aging, generates improvement measures and verifies their effectiveness through digital twin simulation;

[0015] New Project Construction Assistance Module: Extracts geographical environment parameters of new projects, matches cases with a similarity greater than 85% in the historical case library, generates design schemes resistant to salt spray corrosion and turbulence through digital twin simulation, and outputs risk warnings;

[0016] Design and assembly scheme update prompt module: Establish a knowledge graph association between fault types and new schemes, perform accelerated life testing and voiceprint anomaly detection on updated schemes, and generate priority sorting and visual prompts based on fault severity and update cost;

[0017] Maintenance and Management Module: Performs maintenance tasks and records blockchain evidence logs.

[0018] In the fault analysis and tracing module, the input features of the classification model of the random forest model are wind speed fluctuation rate, vibration dominant frequency amplitude, acoustic anomaly index, and environmental salt spray concentration.

[0019] In the fault analysis and tracing module: causal graph nodes are associated with fault phenomena and root cause components, and edge weights represent the strength of causal relationships; the indicators for digital twin simulation verification include vibration reduction rate, temperature stabilization time, and abnormal acoustic signature probability.

[0020] The new project construction support module includes:

[0021] The geographic environment parameter extraction unit analyzes the latitude and longitude, terrain roughness level and annual average salt spray concentration in the planning documents.

[0022] The case matching unit calculates the semantic similarity between the text description of a new project and historical cases using the BERT model;

[0023] The risk-resistance simulation unit injects extreme wind shear and power grid harmonic parameters into the digital twin model to evaluate the design robustness.

[0024] The design and assembly scheme update notification module includes:

[0025] Knowledge graph association unit: Fault type node connects to solution node and applicable model node;

[0026] Accelerated life testing unit: Apply cyclic loads to the new component and monitor whether the stress concentration factor is <1.2.

[0027] The blockchain evidence storage log of the maintenance management module includes:

[0028] Operation timestamp, unit number, maintenance type, replacement part batch number, and maintenance personnel's digital signature;

[0029] Maintenance work order automatic assignment rules: Prioritize matching with the nearest maintenance team whose skill tags match.

[0030] In the data processing module:

[0031] The multi-parameter fusion algorithm adopts a dynamic weighted model, with the following weight allocation: vibration data 40%, acoustic features 30%, stress data 20%, and environmental parameters 10%.

[0032] In the data processing module, the health score threshold is set as follows: ≥80 points is normal, 60-79 points is a warning, and <60 points is an emergency alarm.

[0033] In the data acquisition module: the wind speed and direction sensor is arranged at the top of the tower; the wind turbine blade stress sensing device is a fiber optic grating sensor.

[0034] Compared with the prior art, the beneficial effects of this invention are:

[0035] By combining a random forest classification model with multi-source data (vibration spectrum, acoustic signature features, and environmental parameters), the system can quickly distinguish between internal and external factors of faults, thus avoiding ineffective maintenance caused by misjudgment.

[0036] By constructing a fault root cause map based on cause-effect graph analysis, internal factors such as design defects and manufacturing errors can be located. The effectiveness of improvement measures can be verified by combining digital twin simulation, thereby reducing the recurrence rate of similar faults.

[0037] By matching geographical environmental parameters and semantic analysis of historical case databases, customized design solutions such as anti-turbulence and anti-corrosion are automatically recommended to reduce repetitive design errors.

[0038] Digital twin simulations are used to test design schemes under various scenarios, such as extreme wind conditions and power grid fluctuations (e.g., tower corrosion simulation under salt spray conditions), to identify potential risks in advance and generate reinforcement suggestions, thereby reducing the failure rate of new projects. Historical failure data can be used to guide the construction of new projects (e.g., reusing anti-salt spray coating solutions), forming an experience accumulation and iterative optimization mechanism.

[0039] Based on knowledge graph association technology, validated solutions are automatically matched, avoiding the subjectivity of manual selection. Priority ranking rules are generated according to fault severity, update cost, and expected effect (e.g., formula: Priority = 0.5 × Fault Score + 0.3 × Cost Coefficient + 0.2 × Effect Score).

[0040] This system deeply integrates fault analysis, design optimization, and dynamic updates through a data-driven closed-loop management mechanism, breaking through the limitations of traditional wind farm operation and maintenance systems that rely on human experience. Attached Figure Description

[0041] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0042] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0043] like Figure 1 As shown, a wind farm operation and maintenance system includes a data acquisition module, a data processing module, a fault analysis and tracing module, a new project construction assistance module, and a design and assembly scheme update prompt module.

[0044] The data acquisition module includes fiber optic grating sensors installed at the root of the wind turbine blades, wind speed and direction sensors installed at the top of the tower, a microphone array set up in the nacelle, and environmental salt spray concentration collected by a salt spray concentration sensor.

[0045] Furthermore, the implementation of the fault analysis and tracing module.

[0046] Input maintenance record text (such as abnormal gearbox noise, abnormal oil temperature), and sensor data before the fault occurred (including vibration spectrum, acoustic waveform time-frequency graph, and ambient salt spray concentration).

[0047] The random forest algorithm was adopted, and the input features included wind speed fluctuation rate, vibration dominant frequency amplitude, acoustic anomaly index, and salt spray concentration.

[0048] Construct a fault causal relationship graph in the graph database. Nodes represent fault types (such as bearing wear), components (gearboxes), and design parameters (such as material heat treatment processes). Edges represent causal relationships (such as material defects → stress concentration → crack propagation). Edge weights are obtained from historical fault statistics.

[0049] Establish a digital twin model of the gearbox, input fault parameters (such as tooth surface roughness), simulate the operating state after the improvement scheme (replacing carburized gears), and output vibration amplitude, oil temperature and abnormal sound probability index.

[0050] Compare key parameters before and after the improvement, such as the amplitude of vibration, oil temperature, and probability of abnormal sound patterns in a gearbox.

[0051] Example 1: A turbine at an offshore wind farm experienced motor overheating. The system determined that this was due to salt spray intrusion caused by aging of the sealing ring. Cause-effect diagram analysis traced the root cause to insufficient corrosion resistance of the material, and the corrective measure was to replace the sealing ring with a fluororubber one.

[0052] Furthermore, the implementation of the new project construction support module.

[0053] Extract geographical environmental parameters and generate descriptive text.

[0054] The description of the new project is matched with the project text in the historical case library, and cases with a similarity of >85% are selected as references.

[0055] Import historical fault data of matching cases (such as high tower resonance frequency) and set simulation boundary conditions (such as turbulence intensity).

[0056] Comparing tower reinforcement schemes (increasing wall thickness, adding a constraint damping layer, and changing the cross-sectional shape), simulation results show that adding a constraint damping layer can reduce the resonance probability.

[0057] Example 2: In the initial planning stage of a mountain wind farm, the system matched historical cases (complex terrain caused severe wake effects) and suggested a scheme that combines staggered layout with adaptive yaw.

[0058] Furthermore, the implementation of the design and assembly scheme update prompt module.

[0059] Construct a knowledge graph with nodes including fault type (e.g., loose bolts), solution (e.g., smart gasket), and applicable model.

[0060] Find relevant solutions and prioritize those that have been validated through virtual testing.

[0061] Accelerated life testing was conducted on the updated (verified) scheme to monitor whether its stress concentration factor exceeded the threshold.

[0062] The updated statistics show the unit's trouble-free time and health score.

[0063] Example 3: The turbine units in a wind farm are experiencing excessive vibration due to bearing wear. It is recommended to use bearings with coatings.

[0064] Furthermore, encryption algorithms are employed and encryption keys are dynamically updated; operation logs are recorded and maintained in the blockchain to ensure that the logs are immutable.

[0065] The above description only illustrates preferred embodiments of the present invention, but the present invention is not limited to the above embodiments.

Claims

1. A wind farm operation and maintenance system, characterized in that, include: Data acquisition module: includes wind speed and direction sensors, wind turbine blade stress sensing devices, vibration sensors, microphone arrays and salt spray concentration sensors, used to collect wind conditions, equipment stress, vibration, acoustic data and environmental salt spray concentration in real time; Data processing module: After preprocessing the collected data, it generates a device health score through a multi-parameter fusion algorithm; Fault Analysis and Traceability Module: Based on the random forest model, it distinguishes between internal and external factors of failure, constructs a cause-effect graph for internal factors to trace the root causes of design defects, manufacturing errors or component aging, generates improvement measures and verifies their effectiveness through digital twin simulation; New Project Construction Assistance Module: Extracts geographical environment parameters of new projects, matches cases with a similarity greater than 85% in the historical case library, generates design schemes resistant to salt spray corrosion and turbulence through digital twin simulation, and outputs risk warnings; Design and assembly scheme update prompt module: Establish a knowledge graph association between fault types and new schemes, perform accelerated life testing and voiceprint anomaly detection on updated schemes, and generate priority sorting and visual prompts based on fault severity and update cost; Maintenance and management module: Executes maintenance tasks and records blockchain evidence logs; In the fault analysis and tracing module, the input features of the classification model of the random forest model are wind speed fluctuation rate, vibration dominant frequency amplitude, acoustic anomaly index and environmental salt spray concentration. In the fault analysis and tracing module: the cause-effect graph nodes are associated with the fault phenomena and the root cause components, and the edge weights represent the strength of the causal relationship; The metrics for digital twin simulation verification include vibration descent rate, temperature settling time, and probability of abnormal acoustic signature. The design and assembly scheme update notification module includes: Knowledge graph association unit: Fault type node connects to solution node and applicable model node; Accelerated life testing unit: Apply cyclic loads to the new component and monitor whether the stress concentration factor is <1.2; In the data processing module: The multi-parameter fusion algorithm adopts a dynamic weighted model, with the following weight allocation: vibration data 40%, acoustic features 30%, stress data 20%, and environmental parameters 10%.

2. The wind farm operation and maintenance system according to claim 1, characterized in that: The new project construction support module includes: The geographic environment parameter extraction unit analyzes the latitude and longitude, terrain roughness level and annual average salt spray concentration in the planning documents. The case matching unit calculates the semantic similarity between the text description of a new project and historical cases using the BERT model; The risk-resistance simulation unit injects extreme wind shear and power grid harmonic parameters into the digital twin model to evaluate the design robustness.

3. The wind farm operation and maintenance system according to claim 1, characterized in that: The blockchain evidence storage log of the maintenance management module includes: Operation timestamp, unit number, maintenance type, replacement part batch number, and maintenance personnel's digital signature; Maintenance work order automatic assignment rules: Prioritize matching with the nearest maintenance team whose skill tags match.

4. The wind farm operation and maintenance system according to claim 1, characterized in that: In the data processing module, the health score threshold is set as follows: ≥80 points is normal, 60-79 points is a warning, and <60 points is an emergency alarm.

5. A wind farm operation and maintenance system according to claim 1, characterized in that: In the data acquisition module: the wind speed and direction sensor is arranged at the top of the tower; the wind turbine blade stress sensing device is a fiber optic grating sensor.

Images