Fan vibration active monitoring and control method based on multi-modal sensing and self-adaptive algorithm

Abstract

The invention discloses a multi-modal sensing and self-adaptive algorithm-based fan vibration active monitoring and control method, which comprises the following steps that: 1) a control system fuses fan vibration signals transmitted by different sensors by using a data fusion algorithm to generate vibration data reflecting the running state of a fan; (2) the control system inputs the vibration data and the environment parameters into a machine learning model, and the fan running state is obtained; if the operation state of the fan is abnormal operation, entering the step 3); and (3) the control system adjusts the operation parameters of the fan according to the abnormal type. By integrating the multi-mode sensors (such as an acceleration sensor, a laser displacement sensor and an acoustic sensor) and combining an adaptive algorithm to analyze and process real-time data, the system can dynamically adjust the operation parameters of the fan, active monitoring and suppression of vibration are realized, and the operation stability and reliability of the fan are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of fan vibration monitoring and control, and specifically to an active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms. Background Art

[0002] As the core equipment of modern wind farms, the operating state of fans directly affects power generation efficiency, equipment safety, and operation and maintenance costs. During the operation of a fan, vibration is a common physical phenomenon that may be caused by various reasons, such as the interaction between the blades and the airflow, bearing failures, structural resonances, etc. If these vibrations are not effectively monitored and controlled, it will not only reduce the operating efficiency of the fan, but also may cause fatigue damage to the equipment, increase operation and maintenance costs, and shorten the service life of the equipment.

[0003] Currently, most fan vibration monitoring and operation and maintenance methods mainly rely on a single sensor (such as an acceleration sensor) to collect vibration signals, and judge vibration anomalies by analyzing the frequency or amplitude of the signals. However, this method has the following disadvantages: (1) Single information acquisition: A single sensor can only capture information in a certain dimension of vibration and is difficult to comprehensively reflect the operating state of the fan. For example, the characteristics of displacement and noise may reflect the root cause of vibration, but a single acceleration signal cannot cover these parameters; (2) Lack of real-time control: Existing technologies are mostly passive monitoring, that is, vibration data is recorded and then analyzed offline, and it is impossible to adjust the fan operating parameters in real time to suppress vibration; (3) Insufficient adaptability: The causes of fan vibration are complex and have dynamic characteristics. For example, changes in wind speed and direction, as well as environmental disturbances, will affect vibration, but traditional monitoring methods are difficult to adjust according to changes in external conditions; (4) Lag in vibration analysis: Existing systems usually perform data processing based on fixed algorithms. When facing complex or non-linear vibration modes, it is difficult to accurately identify vibration characteristics, which may lead to false alarms or missed alarms.

[0004] The existence of the above problems is mainly due to the lack of multi-dimensional fusion and intelligent processing capabilities in current technologies for data acquisition and analysis, and at the same time, the complex coupling relationship between the fan operating state and environmental parameters has not been fully considered. Therefore, how to achieve comprehensive monitoring of fan vibration, suppress vibration through real-time control and operation and maintenance strategies, improve the safety and reliability of equipment operation, and reduce operation and maintenance costs has become a research hotspot and difficulty in this field. Summary of the Invention

[0005] The object of the present invention is to provide an active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms, including the following steps:

[0006] 1) Use sensors arranged on the fan to monitor the fan vibration signal and transmit it to the control system;

[0007] Obtain the environmental parameters of the fan and transmit them to the control system

[0008] 2) The control system uses a data fusion algorithm to fuse the fan vibration signals transmitted by different sensors to generate vibration data reflecting the operating state of the fan;

[0009] 3) Construct a machine learning model for optimizing the fan operating parameters;

[0010] 4) The control system inputs the vibration data and environmental parameters into the machine learning model to obtain the operating state of the fan; the operating state of the fan includes normal operation and abnormal operation; when the operating state of the fan is abnormal operation, the operating state of the fan also includes the cause of the abnormality;

[0011] If the operating state of the fan is abnormal operation, then enter step 5);

[0012] 5) The control system adjusts the fan operating parameters according to the type of abnormality.

[0013] Furthermore, the environmental parameters include wind speed, temperature, humidity, and air pressure.

[0014] Furthermore, the sensors include an acceleration sensor, a laser displacement sensor, and an acoustic sensor;

[0015] Among them, the acceleration sensor is installed at the blade and bearing;

[0016] The laser displacement sensor and the acoustic sensor are at least installed on the tower.

[0017] Furthermore, the sensors transmit the fan vibration signals to the control system through a data acquisition module; the data acquisition module transmits signals wirelessly or wired.

[0018] Furthermore, the steps of using a data fusion algorithm to fuse the fan vibration signals transmitted by different sensors include:

[0019] Fuse the fan vibration signals transmitted by the same type of sensors through the weighted average method;

[0020] Fuse the fan vibration signals transmitted by different types of sensors through the Kalman filter algorithm;

[0021] The Kalman filter algorithm is as follows:

[0022]

[0023] Among them, is the estimated state, y k is the measured value, K k is the Kalman gain, and H is the measurement matrix. is the state before estimation.

[0024] The data fusion algorithm includes Kalman filtering and weighted averaging, which are used to process and optimize sensor data.

[0025] Furthermore, the machine learning model includes a neural network and a support vector machine (SVM).

[0026] Furthermore, the machine learning model is trained through a historical data set. The input of the model is vibration data and environmental parameters, and the output is the operating state of the fan.

[0027] Furthermore, during the training process of the machine learning model, a mean square error or cross-entropy loss function is adopted.

[0028] Furthermore, in step 5), the steps for the control system to adjust the fan operating parameters according to the type of abnormality are as follows:

[0029] If the cause of the abnormality is blade imbalance, the control system adjusts the blade angle;

[0030] If the cause of the abnormality is bearing damage, the control system adjusts the fan speed;

[0031] If the cause of the abnormality is environmental factors, the control system adjusts the wind speed of the fan.

[0032] Furthermore, in step 5), after adjusting the fan operating parameters, the control system also generates an operation and maintenance strategy based on the vibration data and historical vibration data.

[0033] The operation and maintenance strategy includes, but is not limited to, a predictive maintenance plan, the operating time and maintenance cycle of the fan, and an operation strategy.

[0034] The technical effect of the present invention is beyond doubt. By integrating multi-modal sensors (such as acceleration sensors, laser displacement sensors, and acoustic sensors) and combining an adaptive algorithm to analyze and process real-time data, the system can dynamically adjust the fan operating parameters, achieve active monitoring and suppression of vibrations, and effectively improve the stability and reliability of the fan operation.

[0035] Specifically, the beneficial effects of the present invention are as follows:

[0036] (1) Multi-dimensional vibration monitoring:

[0037] Fusing the data of multiple sensors to comprehensively obtain the vibration information of the fan, improving the accuracy of monitoring, and providing comprehensive data support for operation and maintenance decisions.

[0038] (2) Real-time control and operation and maintenance optimization:

[0039] Through an adaptive algorithm, the real-time adjustment of the fan operation parameters and operation and maintenance strategies is realized, the abnormal vibration is timely suppressed, the operation and maintenance process is optimized, and the work burden of the operation and maintenance personnel is reduced.

[0040] (3) Improve system adaptability:

[0041] The system can automatically optimize the fan operation parameters and operation and maintenance strategies according to environmental changes, improve the adaptability and stability of the fan, and meet the requirements of flexible operation and maintenance.

[0042] (4) Reduce equipment damage and operation and maintenance costs:

[0043] Through effective vibration suppression and intelligent operation and maintenance strategies, the service life of the equipment is extended, equipment failures caused by vibration are reduced, and maintenance and operation and maintenance costs are lowered.

[0044] Performance improvement:

[0045] (1) Reduction of vibration amplitude: In practical applications, the vibration amplitude of the fan is reduced by about 20%, significantly improving the operation stability.

[0046] (2) Reduction of failure rate: The equipment failure rate is reduced by 15%, extending the average trouble-free time of the equipment.

[0047] (3) Reduction of operation and maintenance costs: Through optimizing the operation and maintenance strategies, the operation and maintenance costs are reduced by 10%, improving the economic benefits. Description of the drawings

[0048] Figure 1 is the flowchart of the method for active monitoring and operation and maintenance control of the fan vibration of the present invention;

[0049] Figure 2 is the control strategy flowchart of the adaptive algorithm of the present invention. Detailed implementation manners

[0050] The present invention will be further described below in conjunction with the embodiments, but it should not be understood that the above-mentioned subject scope of the present invention is limited to the following embodiments. Without departing from the above technical idea of the present invention, various substitutions and modifications made according to the common general knowledge and conventional means in the art should be included within the protection scope of the present invention.

[0051] Embodiment 1:

[0052] See Figures 1 to 2 , the method for active monitoring and control of the fan vibration based on multi-modal sensing and adaptive algorithm includes the following steps:

[0053] 1) Use the sensors arranged on the fan to monitor the fan vibration signals and transmit them to the control system;

[0054] Obtain the environmental parameters of the fan and transmit them to the control system

[0055] 2) The control system uses a data fusion algorithm to fuse the fan vibration signals transmitted by different sensors to generate vibration data reflecting the operating state of the fan;

[0056] 3) Construct a machine learning model for optimizing the operating parameters of the fan;

[0057] 4) The control system inputs the vibration data and environmental parameters into the machine learning model to obtain the operating state of the fan; the operating state of the fan includes normal operation and abnormal operation; when the operating state of the fan is abnormal operation, the operating state of the fan also includes the cause of the abnormality;

[0058] If the operating state of the fan is abnormal operation, then go to step 5);

[0059] 5) The control system adjusts the operating parameters of the fan according to the type of abnormality.

[0060] The environmental parameters include wind speed, temperature, humidity, and air pressure.

[0061] The sensors include an acceleration sensor, a laser displacement sensor, and an acoustic sensor;

[0062] Among them, the acceleration sensor is installed at the blade and bearing;

[0063] The laser displacement sensor and the acoustic sensor are at least installed on the tower.

[0064] The sensors transmit the fan vibration signals to the control system through a data acquisition module; the data acquisition module transmits signals wirelessly or wiredly.

[0065] The steps of fusing the fan vibration signals transmitted by different sensors using a data fusion algorithm include:

[0066] Fuse the fan vibration signals transmitted by the same type of sensors through the weighted average method;

[0067] Fuse the fan vibration signals transmitted by different types of sensors through the Kalman filter algorithm;

[0068] The Kalman filter algorithm is as follows:

[0069]

[0070] Among them, is the estimated state, y k is the measured value, K k is the Kalman gain, and H is the measurement matrix. is the state before estimation.

[0071] The data fusion algorithm includes Kalman filtering and weighted averaging, which are used to process and optimize sensor data.

[0072] The machine learning model includes a neural network and a support vector machine (SVM).

[0073] The machine learning model is trained through a historical dataset. The inputs of the model are vibration data and environmental parameters, and the output is the operating state of the wind turbine.

[0074] During the training process of the machine learning model, a mean square error or cross-entropy loss function is adopted.

[0075] In step 5), the steps for the control system to adjust the operating parameters of the wind turbine according to the type of abnormality are as follows:

[0076] If the cause of the abnormality is blade imbalance, the control system adjusts the blade angle;

[0077] If the cause of the abnormality is bearing damage, the control system adjusts the rotational speed of the wind turbine;

[0078] If the cause of the abnormality is environmental factors, the control system adjusts the wind speed of the wind turbine.

[0079] In step 5), after adjusting the operating parameters of the wind turbine, the control system also generates an operation and maintenance strategy based on the vibration data and historical vibration data.

[0080] The operation and maintenance strategy includes, but is not limited to, a predictive maintenance plan, the operating time and maintenance cycle of the wind turbine, and an operation strategy.

[0081] Embodiment 2:

[0082] A method for active monitoring and control of wind turbine vibration based on multi-modal sensing and adaptive algorithms includes the following steps:

[0083] 1) Use sensors arranged on the wind turbine to monitor the vibration signal of the wind turbine and transmit it to the control system;

[0084] Obtain the environmental parameters where the wind turbine is located and transmit them to the control system.

[0085] 2) The control system uses a data fusion algorithm to fuse the vibration signals of the wind turbine transmitted by different sensors to generate vibration data reflecting the operating state of the wind turbine;

[0086] 3) Construct a machine learning model for optimizing the operating parameters of the wind turbine;

[0087] 4) The control system inputs the vibration data and environmental parameters into the machine learning model to obtain the operating state of the wind turbine. The operating state of the wind turbine includes normal operation and abnormal operation. When the operating state of the wind turbine is abnormal, the operating state also includes the cause of the abnormality;

[0088] If the operating state of the fan is abnormal, go to step 5). If the operating state is normal, return to step 1) and continue monitoring.

[0089] 5) The control system adjusts the fan operating parameters according to the type of abnormality.

[0090] Example 3:

[0091] The active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms has the same technical content as in Example 2. Further, the environmental parameters include wind speed, temperature, humidity, and air pressure.

[0092] Example 4:

[0093] The active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms has the same technical content as any one of Examples 2 - 3. Further, the sensors include acceleration sensors, laser displacement sensors, and acoustic sensors;

[0094] Among them, the acceleration sensors are installed at the blades and bearings;

[0095] The laser displacement sensors and acoustic sensors are installed on the tower and other key parts.

[0096] Example 5:

[0097] The active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms has the same technical content as any one of Examples 2 - 4. Further, the sensors transmit the fan vibration signals to the control system through a data acquisition module; the data acquisition module transmits the signals wirelessly or by wire.

[0098] Example 6:

[0099] The active monitoring and control method for fan vibration based on multi-modal sensing and adaptive algorithms has the same technical content as any one of Examples 2 - 5. Further, the steps of fusing the fan vibration signals transmitted by different sensors using a data fusion algorithm include:

[0100] Fusing the fan vibration signals transmitted by the same type of sensors through the weighted average method;

[0101] Fusing the fan vibration signals transmitted by different types of sensors through the Kalman filter algorithm;

[0102] The Kalman filter algorithm is as follows:

[0103]

[0104] Among them, is the estimated state, y kis the measured value, K k is the Kalman gain, and H is the measurement matrix.

[0105] The data fusion algorithm includes Kalman filtering and weighted averaging for processing and optimizing sensor data.

[0106] Example 7:

[0107] An active monitoring and control method for fan vibration based on multimodal sensing and adaptive algorithms, the technical content is the same as any one of Examples 2-6. Further, the machine learning model includes a neural network and a support vector machine (SVM).

[0108] Example 8:

[0109] An active monitoring and control method for fan vibration based on multimodal sensing and adaptive algorithms, the technical content is the same as any one of Examples 2-7. Further, the machine learning model is trained through a historical data set, the input of the model is vibration data and environmental parameters, and the output is the operating state of the fan.

[0110] Example 9:

[0111] An active monitoring and control method for fan vibration based on multimodal sensing and adaptive algorithms, the technical content is the same as any one of Examples 2-8. Further, during the training process of the machine learning model, the mean square error or cross-entropy loss function is adopted. The selection of the specific loss function depends on the type of output target and the error tolerance.

[0112] Example 10:

[0113] An active monitoring and control method for fan vibration based on multimodal sensing and adaptive algorithms, the technical content is the same as any one of Examples 2-9. Further, in step 5), the steps for the control system to adjust the fan operating parameters according to the abnormal type are:

[0114] If the abnormal reason is blade imbalance, the control system adjusts the blade angle;

[0115] If the abnormal reason is bearing damage, the control system adjusts the fan speed;

[0116] If the abnormal reason is environmental factors, the control system adjusts the wind speed of the fan.

[0117] Example 11:

[0118] An active monitoring and control method for fan vibration based on multimodal sensing and adaptive algorithms, the technical content is the same as any one of Examples 2-10. Further, in step 5), after adjusting the fan operating parameters, the control system also generates an operation and maintenance strategy according to the vibration data and historical vibration data

[0119] The operation and maintenance strategies include, but are not limited to, predictive maintenance plans, the running time and maintenance cycle of the wind turbine, and operation strategies.

[0120] Embodiment 12:

[0121] A method for active monitoring and control of wind turbine vibration based on multimodal sensing and adaptive algorithms, the steps including:

[0122] 1) Use sensors arranged on the wind turbine to monitor the wind turbine vibration signal and transmit it to the control system; obtain the environmental parameters of the wind turbine location and transmit them to the control system, where the environmental parameters include external factors affecting the operation of the wind turbine such as wind speed, temperature, humidity, air pressure, etc.; 2) The control system uses a data fusion algorithm to fuse the wind turbine vibration signals transmitted by different sensors to generate vibration data reflecting the operation state of the wind turbine; for signals between sensors of the same type, such as signals between acceleration sensors, they are fused by the weighted average method; for signals between different types of sensors, such as signals between acceleration sensors, laser displacement sensors, and acoustic sensors, the Kalman filter algorithm is used for fusion, and the formula is as follows:

[0123]

[0124] Where, is the estimated state, y k is the measured value, is K k is the Kalman gain, and H is the measurement matrix.

[0125] 3) Build a machine learning model for optimizing the operation parameters of the wind turbine. The machine learning model can be a neural network, support vector machine (SVM), etc., and is obtained by training with historical data;

[0126] 4) The control system inputs the vibration data and environmental parameters into the machine learning model to obtain the operation parameters of the wind turbine. The output of this machine learning model is the operation state of the wind turbine (such as normal / abnormal vibration). If the output is an abnormal state, the operation parameters of the wind turbine are adjusted according to the type of abnormality.

[0127] For example:

[0128] If the vibration is an abnormality caused by blade imbalance, the control system can adjust the blade angle;

[0129] If the vibration is an abnormality caused by bearing damage, the control system can adjust the wind turbine speed and reduce the step size according to real-time data to avoid further damage.

[0130] If the vibration is an abnormality caused by environmental factors, the wind speed of the wind turbine can be appropriately adjusted to reduce the vibration impact.

[0131] 5) The control system generates operation and maintenance strategies based on the vibration data and historical vibration data. The operation and maintenance strategies include, but are not limited to: predictive maintenance plans to identify in advance the components that may fail and reduce unplanned outages; optimizing the operation time and maintenance cycle of the wind turbine, predicting the failure time of the wind turbine based on historical vibration data, and thus formulating the best maintenance plan; automatically adjusting the operation strategy of the wind turbine according to real-time data to reduce abnormal vibration and improve the operation efficiency of the wind turbine.

[0132] Example 13:

[0133] A method for active monitoring and control of wind turbine vibration based on multi-modal sensing and adaptive algorithms is as follows:

[0134] (1) Multi-sensor fusion monitoring:

[0135] Acceleration sensors, laser displacement sensors and acoustic sensors are installed at key parts of the wind turbine. These sensors respectively obtain different-dimensional data of the wind turbine vibration, and comprehensively analyze the vibration signals from different sensors through data fusion algorithms (such as Kalman filtering, weighted average), so as to comprehensively reflect the operation state of the wind turbine. Compared with the traditional single-sensor scheme, this method provides richer vibration information and provides reliable data support for operation and maintenance decision-making.

[0136] (2) Adaptive algorithm control and operation and maintenance optimization:

[0137] Combined with the adaptive algorithm, aiming at the dynamic characteristics and operation and maintenance requirements of the wind turbine operation, the operation parameters and operation and maintenance strategies of the wind turbine are adjusted in real time. The algorithm can automatically optimize the operation strategy of the wind turbine according to changing factors such as wind speed, wind direction and environmental noise, avoid the lag problem in the traditional method, realize the real-time suppression and control of the wind turbine vibration, and provide real-time optimization suggestions for operation and maintenance personnel. Specifically, the algorithm performs real-time calculations based on the data of different sensors and dynamically adjusts parameters such as the blade angle or rotation speed of the wind turbine, effectively reducing abnormal vibration and extending the service life of the equipment.

[0138] (3) Vibration suppression and operation and maintenance strategy optimization:

[0139] Through real-time monitoring and data analysis, actively identify the abnormal vibration of the wind turbine and quickly adjust the operation parameters for suppression. At the same time, the system generates intelligent operation and maintenance strategies based on vibration characteristics and historical data, such as predictive maintenance plans, to help operation and maintenance personnel formulate scientific maintenance plans, avoid unnecessary outages and repairs, and further reduce operation and maintenance costs.

[0140] Example 14:

[0141] A method for active monitoring and control of wind turbine vibration based on multi-modal sensing and adaptive algorithms is as follows:

[0142] Step 101: Install acceleration sensors, laser displacement sensors, and acoustic sensors at key parts of the fan to monitor the vibration status of the fan respectively.

[0143] Step 102: The data acquisition module transmits the vibration data and environmental data (such as wind speed and wind direction) collected by the sensors to the control system in real time via wireless or wired means.

[0144] Step 103: The control system uses a data fusion algorithm to fuse the data obtained from different sensors in real time to generate a comprehensive vibration signal, reflecting the overall operating status of the fan.

[0145] Step 104: The adaptive algorithm analyzes and optimizes the operating parameters of the fan in real time according to the fused vibration data and external environmental changes, such as adjusting the blade angle and changing the rotational speed.

[0146] Step 105: The system automatically selects the most appropriate control strategy according to the vibration characteristics to suppress abnormal vibrations and generates corresponding operation and maintenance strategies, such as preventive maintenance plans, for the operation and maintenance personnel to refer to.

[0147] Step 106: Apply the optimized operating parameters and operation and maintenance strategies to the control system of the fan to achieve active monitoring and operation and maintenance control of the fan vibration.

[0148] Example 15:

[0149] An active fan vibration monitoring and control system based on multi-modal sensing and adaptive algorithms, including a fan vibration monitoring system and an operation and maintenance control module.

[0150] Vibration monitoring system: Composed of acceleration sensors, laser displacement sensors, and acoustic sensors, installed at key parts such as the blades, bearings, and tower of the fan to collect vibration data in real time.

[0151] Data fusion module: Adopt the Kalman filter algorithm to fuse and process multi-sensor data to obtain high-precision vibration status information.

[0152] Adaptive algorithm module: Based on machine learning models (such as neural networks and support vector machines), analyze vibration data and environmental parameters in real time to optimize the operating parameters and operation and maintenance strategies of the fan.

[0153] Operation and maintenance control module: Adjust the operating status of the fan according to the output of the adaptive algorithm and provide maintenance suggestions to the operation and maintenance personnel.

[0154] In the actual application of a wind farm, the system of the present invention successfully monitored the abnormal vibration of a wind turbine. Through the adaptive algorithm, the system adjusted the blade angle of the wind turbine in real time, suppressing the abnormal vibration. At the same time, the system generated preventive maintenance suggestions, prompting the operation and maintenance personnel to check the relevant components, ultimately avoiding a possible equipment failure and reducing the downtime and maintenance costs.

Claims

1. A method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm, characterized in that: The following steps are involved: 1) Use sensors arranged on the fan to monitor the fan vibration signal and transmit it to the control system; Obtain the environmental parameters of the fan and transmit them to the control system 2) The control system uses a data fusion algorithm to fuse the fan vibration signals transmitted by different sensors to generate vibration data reflecting the operating status of the fan; 3) Build a machine learning model for optimizing wind turbine operating parameters; 4) The control system inputs the vibration data and environmental parameters into the machine learning model to obtain the fan operation status; the fan operation status includes normal operation and abnormal operation; when the fan operation status is abnormal operation, the fan operation status also includes the abnormal reason; If the fan operation status is abnormal, go to step 5); 5) The control system adjusts the fan operating parameters according to the type of abnormality.

2. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1 is characterized in that: The environmental parameters include wind speed, temperature, humidity, and air pressure.

3. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1 is characterized in that: The sensors include acceleration sensors, laser displacement sensors and acoustic sensors; Wherein, the acceleration sensor is installed at the blade and the bearing; The laser displacement sensor and the acoustic sensor are at least installed on the tower.

4. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1 is characterized in that: The sensor transmits the fan vibration signal to the control system through the data acquisition module; the data acquisition module transmits the signal wirelessly or wiredly.

5. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1 is characterized in that: The steps of fusing the fan vibration signals transmitted by different sensors using the data fusion algorithm include: The fan vibration signals transmitted by sensors of the same type are fused by weighted average method; The fan vibration signals transmitted by different types of sensors are fused through the Kalman filter algorithm; The Kalman filter algorithm is as follows: in, is the estimated state, y k is the measured value, K k is the Kalman gain, H is the measurement matrix; is the state before estimation. The data fusion algorithm includes Kalman filtering and weighted averaging, which are used to process and optimize sensor data.

6. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1 is characterized in that: The machine learning model includes a neural network and a support vector machine (SVM).

7. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 6 is characterized in that: The machine learning model is trained using historical data sets, the input of the model is vibration data and environmental parameters, and the output is the operating status of the fan.

8. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 7 is characterized in that: During the training process of the machine learning model, a mean square error or a cross entropy loss function is used.

9. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1, characterized in that: In step 5), the control system adjusts the fan operating parameters according to the abnormality type as follows: If the abnormality is caused by blade imbalance, the control system adjusts the blade angle; If the abnormality is caused by bearing damage, the control system adjusts the fan speed; If the abnormality is caused by environmental factors, the control system adjusts the wind speed of the fan.

10. The method for active monitoring and control of fan vibration based on multimodal sensing and adaptive algorithm according to claim 1, characterized in that: In step 5), after adjusting the fan operating parameters, the control system also generates an operation and maintenance strategy based on the vibration data and historical vibration data. The operation and maintenance strategy includes but is not limited to predictive maintenance plans, wind turbine operating hours and maintenance cycles, and operation strategies.

Images