Fault early warning system for gear box and generator in wind generating set

Abstract

The invention provides a fault early warning system for a gearbox and a generator in a wind generating set. The fault early warning system comprises an intelligent sound acquisition device, a data transmission device, a background analysis device and a central control room front-end display device which are connected in sequence, the intelligent sound collection device comprises a sound listening device, a sound collector, a sound insulation part and a sound processing chip. One end of the listening device is fixed between the generator and the gearbox, the other end of the listening device is connected with the sound collector, the connecting part of the listening device and the sound collector is wrapped by the sound insulation component, and the sound collector is connected with the sound processing chip through a wire; and the data transmission device transmits the compressed and packaged data to the background analysis device, analyzes the data and transmits an analysis result to the front-end display device of the central control room to display the analysis result. According to the invention, the technical problems that short-time inspection cannot be realized, the accuracy of fault discrimination by sound cannot be ensured and safety risks exist due to the adoption of the technical scheme of fault discrimination by sound listening of a thermal power plant in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of wind power generation, and more particularly to a fault early warning system for the gearbox and generator in a wind turbine generator set. Background Technology

[0002] In the current wind power industry, generator and gearbox monitoring is mainly based on temperature. However, when the generator winding, bearing temperature and gearbox oil temperature exceed the limit, the equipment is basically in a serious fault state. At this time, maintenance and repair require spare parts and the grid cannot be connected for a long time, resulting in damage to the wind farm's benefits.

[0003] In thermal power plants, experienced employees can use listening devices to determine if there are potential faults in the gearbox and generator's internal operating equipment. This allows them to detect equipment defects in advance, carry out planned maintenance, and prepare spare parts in advance.

[0004] However, this method is not applicable in wind farms. First, while equipment in thermal power plants is relatively concentrated and can be manually inspected in a short time, wind turbines in wind farms are more dispersed and cannot be inspected in a short time. Second, there are fewer maintenance personnel in wind farms compared to thermal power plants, making it impossible to ensure the accuracy of identifying faults by sound. Third, inspecting gearboxes and generators requires climbing to the top of the nacelle and hearing the operating sound only when the unit is running, which poses a significant safety risk. Summary of the Invention

[0005] Based on the above problems, this invention proposes a fault early warning system for the gearbox and generator within a wind turbine generator set. This invention solves the technical problems inherent in existing technologies that rely on sound-based fault identification, such as the inability to conduct timely inspections, the inability to ensure the accuracy of sound-based fault identification, and the significant safety risks. This invention integrates a smart sound acquisition device, a data transmission device, a backend analysis device, and a central control room front-end display device into a single system, enabling wind turbine generator sets to also utilize sound-based fault identification with very high accuracy.

[0006] This invention proposes a fault early warning system for the gearbox and generator within a wind turbine generator set, comprising: The system consists of a sound intelligent acquisition device, a data transmission device, a background analysis and resolution device, and a central control room front-end display device, connected in sequence. The intelligent sound acquisition device includes: a listener, a sound acquisition unit, sound insulation components, and a sound processing chip; One end of the listening device is fixed between the generator and the gearbox, and the other end of the listening device is connected to the sound acquisition device. The connection between the listening device and the sound acquisition device is wrapped with soundproofing components. The sound acquisition device and the sound processing chip are connected by wires. The listening device transmits the sound from inside the generator and gearbox to the external sound acquisition device. The sound acquisition device transmits the signal to the sound processing chip. The sound processing chip preprocesses and compresses the received sound signal and transmits the compressed and packaged data to the data transmission device. The data transmission device transmits the compressed and packaged data to the background parsing and analysis device. The background parsing and analysis device analyzes the data and transmits the analysis results to the front-end display device in the central control room, which then displays the analysis results.

[0007] In addition, the data transmission device includes: a first photoelectric conversion module, an optical fiber, and a second photoelectric conversion module; The first photoelectric conversion module receives the compressed and packaged data and converts the electrical signal data into optical signal data. The optical signal data is transmitted to the background analysis device through optical fiber. The second photoelectric conversion module, which is located in the background analysis device, converts the optical signal data into electrical signal data.

[0008] In addition, the background analysis device is equipped with a background server, which contains analysis algorithm software. After receiving the electrical signal data, the back-end server first decompresses it, analyzes the decompressed data using analysis algorithm software, and transmits the location of the fault in the generator and / or gearbox to the front-end display device in the central control room.

[0009] In addition, the analysis algorithm software collects sound data from multiple generators and gearboxes in advance, marks suspicious sounds through manual analysis and marking, verifies on-site whether the suspicious equipment corresponding to the suspicious sounds has potential faults, and records the verification results to form an algorithm database. When the analysis algorithm software analyzes the decompressed data, it compares the data with the data in the algorithm database to determine the location of the fault in the generator and gearbox.

[0010] In addition, the algorithm model is trained using data from the algorithm database. The algorithm model identifies the decompressed data, automatically locates the fault location, generates maintenance suggestions, and synchronizes the maintenance suggestions to the external operation and maintenance management system.

[0011] In addition, the central control room's front-end display device is equipped with an industrial control computer and monitoring software. The industrial control computer is used to receive the analysis results sent by the background analysis device, and the monitoring software is used to display and issue control commands.

[0012] In addition, one end of the listening device is fixed between the generator and the gearbox via a magnetic base.

[0013] In addition, the listening device is either a listening stick or a magnetic high-sensitivity sound sensor.

[0014] In addition, temperature and humidity sensors and vibration sensors are integrated next to the sound intelligent acquisition device to simultaneously collect temperature and vibration data of the gearbox and generator.

[0015] In addition, the intelligent sound acquisition device also includes an audible and visual alarm connected to the sound processing chip. The sound processing chip is embedded with a simple fault identification algorithm. When the simple fault identification algorithm detects that the acquired sound frequency exceeds a preset threshold, it triggers the audible and visual alarm.

[0016] This invention solves the technical problems of existing technologies that rely on sound-based fault identification in thermal power plants, such as the inability to conduct short-term inspections, the inability to ensure the accuracy of sound-based fault identification, and the significant safety risks. This invention integrates a smart sound acquisition device, a data transmission device, a back-end analysis device, and a central control room front-end display device into a single system, enabling wind turbine generators to also employ sound-based fault identification with very high accuracy. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a fault early warning system for the gearbox and generator in a wind turbine generator set, provided as an embodiment of the present invention. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to specific embodiments and accompanying drawings. This description is intended only to illustrate specific embodiments of the invention and does not constitute any limitation on the invention. The scope of protection of the invention is defined by the claims.

[0019] Reference Figure 1 This invention proposes a fault early warning system for the gearbox and generator within a wind turbine generator set, comprising: The following components are connected in sequence: intelligent sound acquisition device 1, data transmission device 2, background analysis device 3, and central control room front-end display device 4. The intelligent sound acquisition device 1 includes: a listener 11, a sound acquisition unit 12, a sound insulation component 13, and a sound processing chip 14; One end of the listening device 11 is fixed between the generator and the gearbox 5, and the other end of the listening device 11 is connected to the sound acquisition device 12. The connection between the listening device 11 and the sound acquisition device 12 is wrapped with a soundproof component 13. The sound acquisition device 12 is connected to the sound processing chip 14 through a wire. The listening device 11 transmits the sound inside the generator and gearbox 5 to the external sound collector 12. The sound collector 12 transmits the signal to the sound processing chip 14. The sound processing chip 14 preprocesses and compresses the received sound signal and transmits the compressed and packaged data to the data transmission device 2. Data transmission device 2 transmits the compressed and packaged data to background parsing and analysis device 3. Background parsing and analysis device 3 analyzes the data and transmits the analysis results to the front-end display device 4 in the central control room. The front-end display device 4 in the central control room displays the analysis results.

[0020] First, one end of the earpiece 11 of the intelligent sound acquisition device 1 is fixed at a suitable position between the generator and the gearbox 5. The other end of the earpiece 11 is connected to the sound acquisition unit 12. The earpiece 11 can transmit the sound inside the generator and gearbox to the outside. A soundproof component 13 is used to enclose the connection between the earpiece 11 and the sound acquisition unit 12 to prevent external sound interference with sound acquisition and transmission. The sound acquisition unit 12 transmits the signal to the sound processing chip 14. The sound processing chip 14 preprocesses and compresses the received sound signal to improve the subsequent transmission rate.

[0021] By fixing the listening device 11 to a specific position between the generator and the gearbox 5, efficient transmission of mechanical vibration sound is achieved; at the same time, a sound insulation component 13 is used to wrap the connection between the listening device 11 and the sound acquisition device 12. Optionally, the sound insulation component 13 is sound insulation cotton. The sound insulation component 13 is used to shield external environmental noise interference to ensure the accuracy and stability of the acquired data.

[0022] Alternatively, the listening device can be a listening stick.

[0023] The sound acquisition unit 12 transmits the acquired raw signal to the sound processing chip, which performs real-time preprocessing and compression packaging, effectively reducing data volume, improving the efficiency of subsequent fiber optic transmission, reducing transmission delay, and enhancing the timeliness of fault early warning.

[0024] Optionally, the data transmission device 2 includes: a first photoelectric conversion module 21, an optical fiber 22, and a second photoelectric conversion module 23; The first photoelectric conversion module 21 receives the compressed and packaged data and converts the electrical signal data into optical signal data. The optical signal data is transmitted to the background analysis and interpretation device 3 through optical fiber. The second photoelectric conversion module 23, which is set in the background analysis and interpretation device 3, converts the optical signal data into electrical signal data.

[0025] By combining photoelectric conversion modules with fiber optic transmission, long-distance, highly reliable data transmission is achieved, avoiding signal attenuation and interference problems, thus realizing stable long-distance audio signal transmission based on photoelectric conversion modules.

[0026] The use of dual photoelectric conversion modules ensures the anti-interference capability and signal integrity of sound signals during long-distance transmission (such as from the wind farm site to the back-end server), significantly improving stability and reliability in industrial environments.

[0027] Optionally, the background analysis device 3 is equipped with a background server 31, and the background server 31 is equipped with analysis algorithm software. After receiving the electrical signal data, the backend server 31 first decompresses the data, analyzes the decompressed data using analysis algorithm software, and transmits the location of the fault in the generator and / or gearbox to the front-end display device 4 in the central control room.

[0028] Optionally, the front-end display device 4 in the central control room is equipped with an industrial control computer 41 and monitoring software. The industrial control computer 41 is used to receive the analysis results sent by the background analysis device 3, and the monitoring software is used to display and issue control commands.

[0029] This invention solves the technical problems of existing technologies that rely on sound-based fault identification in thermal power plants, such as the inability to conduct short-term inspections, the inability to ensure the accuracy of sound-based fault identification, and the significant safety risks. This invention integrates a smart sound acquisition device, a data transmission device, a back-end analysis device, and a central control room front-end display device into a single system, enabling wind turbine generators to also employ sound-based fault identification with very high accuracy.

[0030] This invention provides a fault early warning system for the gearbox and generator within a wind turbine generator set. By efficiently collecting, stably transmitting, and intelligently analyzing the sound signals generated during equipment operation, it achieves early identification and precise location of potential equipment faults, improving the operation and maintenance efficiency and safety of the wind turbine generator set, and reducing equipment failure rates and maintenance costs. Utilizing the mechanical conduction characteristics of the audio receiver, the sound signals from inside the gearbox and generator are stably transmitted to the sound acquisition unit. Combined with sound insulation components, external environmental interference is effectively shielded, ensuring the purity and accuracy of the collected data. A local sound processing chip preprocesses and compresses the raw sound signals, and combined with a photoelectric conversion module and fiber optic transmission, long-distance, high-reliability data transmission is achieved, avoiding signal attenuation and interference problems. The background analysis algorithm software automatically identifies sound features and continuously optimizes the algorithm using manually labeled training samples, accurately predicting potential faults in the generator and gearbox, enabling early warning and assisting maintenance personnel in timely troubleshooting of equipment hazards. The analysis results are transmitted to the front-end display device in the central control room in real time. The front-end monitoring software can intuitively display the equipment status and can cooperate with the control system to issue operation and maintenance instructions, realizing closed-loop management from anomaly detection and alarm prompts to fault handling, thereby improving the overall intelligent operation and maintenance level of the wind farm.

[0031] In one embodiment, the data transmission device 2 includes: a first photoelectric conversion module 21, an optical fiber 22, and a second photoelectric conversion module 23; The first photoelectric conversion module 21 receives the compressed and packaged data and converts the electrical signal data into optical signal data. The optical signal data is transmitted to the background analysis and interpretation device 3 through optical fiber. The second photoelectric conversion module 23, which is set in the background analysis and interpretation device 3, converts the optical signal data into electrical signal data.

[0032] By combining photoelectric conversion modules with fiber optic transmission, long-distance, highly reliable data transmission is achieved, avoiding signal attenuation and interference problems, thus realizing stable long-distance audio signal transmission based on photoelectric conversion modules.

[0033] The use of dual photoelectric conversion modules ensures the anti-interference capability and signal integrity of sound signals during long-distance transmission (such as from the wind farm site to the back-end server), significantly improving stability and reliability in industrial environments.

[0034] In one embodiment, the background parsing and analysis device 3 is equipped with a background server 31, and the background server 31 is equipped with analysis algorithm software. After receiving the electrical signal data, the backend server 31 first decompresses the data, analyzes the decompressed data using analysis algorithm software, and transmits the location of the fault in the generator and / or gearbox to the front-end display device 4 in the central control room.

[0035] The background server 31 is mainly used to identify whether the sound signal is a fault sound, and then determine the location of the fault based on the identification result.

[0036] In one embodiment, the analysis algorithm software collects sound data from multiple generators and gearboxes in advance, marks suspicious sounds through manual analysis and labeling, verifies on-site whether the suspicious equipment corresponding to the suspicious sounds has potential faults, and records the verification results to form an algorithm database. When the analysis algorithm software analyzes the decompressed data, it compares the data with the data in the algorithm database to determine the location of the fault in the generator and gearbox.

[0037] Optionally, to achieve a precise correspondence between "suspicious sound tagging" and "equipment malfunction verification," a correspondence system of "sound data - collection scenario - equipment status" is established, using a unique identifier throughout the entire process. This can be divided into three key steps: 1. Give each piece of audio data a unique identifier. When the intelligent sound acquisition device collects data, it automatically generates a unique ID for each sound segment (such as a 10-second audio file). The tag must contain three types of core information: Equipment identification: Fan number + generator / gearbox number (e.g., "Fan 08-Gearbox A"); Data collection scenarios: data collection time (accurate to the second), data collection location (fixed point of the listener, such as "the bearing at the input end of the gearbox"), environmental parameters (wind speed and equipment load rate at the time); Data attributes: sound file format, sampling rate, and path to the corresponding original electrical signal file.

[0038] These tags are compressed and packaged along with the sound data, and then transmitted to the backend and bound to the audio file for storage, ensuring that each sound can be traced back to the specific device and collection scenario. 2. Manual markings and on-site verification are matched one-to-one. During the backend analysis, after manually filtering out suspicious sounds (such as abnormal frequencies or background noise), a "verification work order" is generated based on the "unique identifier" from the first step. The work order clearly states: Equipment requiring verification: Locate the specific component of the specific fan based on the "Equipment Identification"; Scenario requiring verification: Based on the "collection time + location", maintenance personnel can collect the sound again under the same operating conditions (such as similar wind speed and load) to compare whether the suspicious signal is reproduced; Verification content: Clarify the type of fault to be checked (e.g., "determine if there is bearing wear in the gearbox").

[0039] After on-site verification, maintenance personnel fill in the results such as "whether there is a fault" and "fault type" into the work order, and then use the system to associate the corresponding suspicious sound tag to complete the "tag-verification" binding.

[0040] 3. Relationships are preserved during algorithm iteration, and labels are continuously corrected. During the correction phase, the system will automatically record the matching results between each "algorithm judgment result" and "on-site verification result": If the algorithm determines that a certain sound is "suspicious", but on-site verification shows no fault, the "label" of that sound will be updated to "false positive sample" to optimize the algorithm's recognition threshold; If the algorithm does not determine that a problem is suspected, but a fault is found on-site, the "label" of the corresponding sound will be updated to "missed sample" and added to the algorithm's training set.

[0041] In this way, each correction can be precisely mapped to the original sound data, ensuring that the algorithm continues to iterate towards "high accuracy".

[0042] By collecting a large number of sound samples from actual operation, and combining manual annotation with on-site verification, a sound feature recognition algorithm is gradually optimized, ultimately achieving a fault diagnosis accuracy rate of over 95%, and enabling accurate prediction of potential fault hazards in gearboxes and generators.

[0043] In one embodiment, the algorithm model is trained using data from the algorithm database. The algorithm model identifies the decompressed data, automatically locates the fault location, generates maintenance suggestions, and synchronizes the maintenance suggestions to the external operation and maintenance management system.

[0044] By automatically locating the fault location and generating maintenance suggestions, the maintenance suggestions are synchronized with the external operation and maintenance management system, making the whole process more intelligent.

[0045] In one embodiment, the central control room front-end display device 4 is equipped with an industrial control computer 41 and monitoring software. The industrial control computer 41 is used to receive the analysis results sent by the background parsing and analysis device 3, and the monitoring software is used to display and issue control commands.

[0046] Through the linkage between the back-end server 31 and the industrial control computer 41, the analysis results are displayed in real time through the monitoring software. Combined with the issuance of control commands, the entire closed-loop control process from sound anomaly detection to alarm and command response is realized, thereby improving the operational safety of wind turbine units.

[0047] In one embodiment, one end of the listening device 11 is fixed between the generator and the gearbox by a magnetic base.

[0048] The magnetic attachment allows this system to be easily moved to a different location.

[0049] In one embodiment, the listener 11 is a listening stick or a magnetic high-sensitivity acoustic sensor.

[0050] The listening stick can accurately capture faint sounds: using the principle of conduction, it can collect faint sounds emitted by the equipment during operation. These sounds cannot be heard directly by the ear due to air isolation and external interference, just like a doctor's stethoscope, which can accurately determine the operating status of the equipment.

[0051] The advantage of using a magnetic high-sensitivity acoustic sensor is that it has high sensitivity and wide frequency response, and can capture minute changes inside the material, such as cracks or stress concentrations.

[0052] In one embodiment, a temperature and humidity sensor and a vibration sensor are integrated next to the sound intelligent acquisition device 1 to simultaneously collect temperature and vibration data of the gearbox and generator.

[0053] By monitoring the generator and gearbox of the wind turbine together with temperature or vibration data, a more comprehensive monitoring and fault diagnosis can be achieved.

[0054] In one embodiment, the sound intelligent acquisition device 1 also includes an audible and visual alarm connected to a sound processing chip. A simple fault identification algorithm is embedded in the sound processing chip. When the simple fault identification algorithm detects that the acquired sound frequency exceeds a preset threshold, it triggers the audible and visual alarm.

[0055] The preset threshold is set based on experience, for example, 500Hz. By promptly identifying simple faults, an alarm can be issued in a timely manner to prevent equipment damage.

[0056] As needed, the above technical solutions can be combined to achieve the best technical effect.

[0057] The above description is merely the principle and preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several other modifications can be made based on the principle of the present invention, and these modifications should also be considered within the scope of protection of the present invention.

Claims

1. A fault early warning system for the gearbox and generator within a wind turbine generator set, characterized in that, include: The system consists of a sound intelligent acquisition device, a data transmission device, a background analysis and resolution device, and a central control room front-end display device, connected in sequence. The intelligent sound acquisition device includes: a listener, a sound acquisition unit, sound insulation components, and a sound processing chip; One end of the listening device is fixed between the generator and the gearbox, and the other end of the listening device is connected to the sound acquisition device. The connection between the listening device and the sound acquisition device is wrapped with soundproofing components. The sound acquisition device and the sound processing chip are connected by wires. The listening device transmits the sound from inside the generator and gearbox to the external sound acquisition device. The sound acquisition device transmits the signal to the sound processing chip. The sound processing chip preprocesses and compresses the received sound signal and transmits the compressed and packaged data to the data transmission device. The data transmission device transmits the compressed and packaged data to the background parsing and analysis device. The background parsing and analysis device analyzes the data and transmits the analysis results to the front-end display device in the central control room, which then displays the analysis results.

2. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 1, characterized in that, The data transmission device includes: a first photoelectric conversion module, an optical fiber, and a second photoelectric conversion module; The first photoelectric conversion module receives the compressed and packaged data and converts the electrical signal data into optical signal data. The optical signal data is transmitted to the background analysis device through optical fiber. The second photoelectric conversion module, which is located in the background analysis device, converts the optical signal data into electrical signal data.

3. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 2, characterized in that, The background analysis device is equipped with a background server, and the background server has analysis algorithm software installed. After receiving the electrical signal data, the back-end server first decompresses it, analyzes the decompressed data using analysis algorithm software, and transmits the location of the fault in the generator and / or gearbox to the front-end display device in the central control room.

4. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 3, characterized in that, The analysis algorithm software collects sound data from multiple generators and gearboxes in advance, marks suspicious sounds through manual analysis and labeling, verifies on-site whether the suspicious equipment corresponding to the suspicious sounds has potential faults, and records the verification results to form an algorithm database. When the analysis algorithm software analyzes the decompressed data, it compares the data with the data in the algorithm database to determine the location of the fault in the generator and gearbox.

5. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 4, characterized in that, The algorithm model is trained using data from the algorithm database. The algorithm model identifies the decompressed data, automatically locates the fault location, generates maintenance suggestions, and synchronizes the maintenance suggestions to the external operation and maintenance management system.

6. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 1, characterized in that, The central control room's front-end display device is equipped with an industrial control computer and monitoring software. The industrial control computer is used to receive the analysis results sent by the background analysis device, while the monitoring software is used to display and issue control commands.

7. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 1, characterized in that, One end of the audio receiver is fixed between the generator and the gearbox via a magnetic base.

8. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 1, characterized in that, The listening device is either a listening stick or a magnetic high-sensitivity sound sensor.

9. The fault early warning system for the gearbox and generator in a wind turbine generator set according to claim 1, characterized in that, Temperature and humidity sensors and vibration sensors are integrated next to the sound intelligent acquisition device to simultaneously collect temperature and vibration data of the gearbox and generator.

10. The fault early warning system for the gearbox and generator in a wind turbine generator set according to any one of claims 1-9, characterized in that, The sound intelligent acquisition device also includes an audible and visual alarm connected to a sound processing chip. A simple fault identification algorithm is embedded in the sound processing chip. When the simple fault identification algorithm detects that the acquired sound frequency exceeds a preset threshold, the audible and visual alarm is triggered.

Images