System for reducing shaft current damage of generator and working method

Abstract

The invention relates to the technical field of wind driven generators, in particular to a system for reducing generator shaft current damage and a working method.The system is characterized in that a generator spindle is connected with a generator shell through at least one insulating bearing, and at least one grounding ring is further arranged on the generator spindle; the grounding ring is grounded through a carbon brush and a grounding wire which are connected in sequence; the grounding wire is provided with a current transformer, and the current transformer is used for obtaining grounding current data and transmitting the grounding current data to the data processing end. And the data processing end is used for predicting working state data of the insulating bearing based on the grounding current data, the data of the environment where the fan is located, the fan operation data and the insulating bearing data, and evaluating whether a shaft current risk alarm is sent or not based on the working state data of the insulating bearing. By monitoring and evaluating the working state of the insulating bearing of the system, operation and maintenance personnel can master the working state of the insulating material of the insulating bearing at any time, and the situation that equipment is damaged by shaft current due to failure of the insulating material is avoided.

Description

Technical Field

[0001] The invention belongs to the technical field of wind turbines, and specifically discloses a system and a working method for reducing shaft current damage of a generator. Background Art

[0002] During operation, wind turbines generate shaft currents due to factors such as magnetic circuit asymmetry. Large shaft currents can cause irreversible damage to the equipment. Shaft currents are caused by shaft voltage and can be categorized into two types based on the path they take: differential-mode shaft current and common-mode shaft current. Differential-mode shaft current is the current flowing through the shaft, bearings, casing, other bearing, and finally the shaft; common-mode shaft current is the current flowing through the shaft, bearings at both ends, casing, and ground.

[0003] The emergence of shaft current causes premature failure of bearings, resulting in erosion of bearing metal materials and degradation of lubricant functions. It may also cause vibration, shutdown, and burnout of the entire equipment, seriously affecting the safe operation of electrical equipment and increasing operating costs. Therefore, suppressing shaft current is an important way to reduce the cost of wind power generation. Summary of the Invention

[0004] The purpose of the present invention is to provide a system and working method for reducing generator shaft current damage, so as to achieve stable suppression of shaft current; the specific scheme is as follows:

[0005] In a first aspect, a system for reducing generator shaft current damage is provided, wherein a generator main shaft is connected to a generator housing via at least one insulating bearing, and at least one grounding ring is further provided on the generator main shaft;

[0006] The grounding ring is grounded by sequentially connected carbon brushes and grounding wires;

[0007] A current transformer is provided on the grounding wire, and the current transformer is used to obtain grounding current data and transmit it to the data processing end;

[0008] The data processing end is used to predict the insulating bearing working status data based on the ground current data, the wind turbine environment data, the wind turbine operation data and the insulating bearing data, and to evaluate whether to issue a shaft current risk alarm based on the insulating bearing working status data.

[0009] Furthermore, there are two insulating bearings, and the driving end and the non-driving end of the generator are both connected to the generator housing through the insulating bearings.

[0010] Furthermore, there are two grounding rings, and both the driving end and the non-driving end of the generator are provided with the grounding ring.

[0011] Furthermore, it also includes a sensor module, which includes a temperature and humidity sensor, a salt spray concentration sensor, and a vibration sensor; the temperature and humidity sensor, the salt spray concentration sensor, and the vibration sensor are all connected to the data processing end.

[0012] Furthermore, the grounding current data includes current value; the environmental data of the fan includes temperature value, humidity value and salt spray concentration value; the fan operation data includes vibration signal data; and the insulating bearing data includes insulation layer thickness, dielectric strength and material aging coefficient.

[0013] Furthermore, the data processing end has a built-in insulating bearing working state prediction model, which obtains the current working state of the insulating bearing by importing the real-time current value, temperature value, humidity value, salt spray concentration value, and vibration signal data into the insulating bearing working state prediction model.

[0014] Furthermore, the working status assessment is that when the working status data of the insulating bearing is within a preset insulating bearing working status range, it is assessed as being in good condition; otherwise, it is assessed as being in poor condition, and a shaft current risk alarm is activated.

[0015] In a second aspect, a method for reducing generator shaft current damage is provided, which is used in the above-mentioned system for reducing generator shaft current damage, and includes:

[0016] Real-time acquisition of ground current data, wind turbine environment data, and wind turbine operation data and pre-processing;

[0017] The pre-processed data is input into the insulated bearing working state prediction model to obtain the current insulated bearing working state;

[0018] Based on the current insulated bearing working status data and the preset insulated bearing working status range, it is evaluated whether to issue a shaft current risk alarm.

[0019] Furthermore, the preprocessing includes:

[0020] Perform outlier cleaning and interpolation processing on current values, temperature values, humidity values, salt spray concentration values and vibration signal data;

[0021] The current fluctuation variance data is obtained based on the processed current value; the average temperature, average humidity and average salt spray concentration in the data collection period are obtained based on the temperature, humidity and salt spray concentration values; the frequency domain characteristics of the vibration signal are obtained based on the processed vibration signal data.

[0022] Furthermore, the insulating bearing working state prediction model is constructed using a two-layer fully connected neural network;

[0023] The dimension of the first layer of the neural network is 64 dimensions, and the expression is as follows:

[0024] h1=ReLU(W1x+b1);

[0025] The dimension of the second layer of the neural network is 32 dimensions, and the expression is as follows:

[0026] h2=ReLU(W2h1+b2);

[0027] The dimension of the output layer is 1, and the expression is as follows:

[0028] y=σ(W3h2+b3);

[0029] Among them, x is the input matrix, W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias vectors, and ReLU is the activation function.

[0030] Beneficial effects of the present invention:

[0031] The present invention uses insulating bearings and insulating end caps to change the impedance of the shaft current path, blocking the shaft current path. By grounding the shaft at the generator's drive and / or non-drive ends, a bypass is created for the shaft current, reducing the shaft current flowing through the bearings. Current transformers are installed on the shaft grounding cables at the generator's drive and / or non-drive ends to monitor the shaft current. The detected values are sent to a data processing terminal for analysis and processing to form an assessment of the insulated bearing's operating status. This guides timely replacement of the bearing's insulation material, reduces operation and maintenance costs, and ensures equipment safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a diagram of the equipment structure of the system for reducing generator shaft current damage provided by the present invention;

[0033] Figure 2 A data flow diagram of a system for reducing generator shaft current damage provided by the present invention;

[0034] Figure 3 This is a flow chart of the working method for reducing generator shaft current damage provided by the present invention.

[0035] Figure markings: 1-generator housing, 2-insulating end cover of main shaft drive end, 3-grounding ring of main shaft drive end, 4-carbon brush of main shaft drive end, 5-current transformer of main shaft drive end, 6-grounding wire of main shaft drive end, 7-insulating bearing of main shaft drive end, 8-insulating end cover of main shaft non-drive end, 9-grounding ring of main shaft non-drive end, 10-carbon brush of main shaft non-drive end, 11-current transformer of main shaft non-drive end, 12-grounding wire of main shaft non-drive end, 13-insulating bearing of main shaft non-drive end. DETAILED DESCRIPTION

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0037] Wind turbines are expensive, with the main shaft bearings accounting for approximately 2% of the total cost. The presence of shaft currents can cause premature bearing failure, leading to metal erosion and lubricant degradation, and potentially causing vibration, shutdown, and burnout of the entire unit, seriously impacting the safe operation of electrical equipment and increasing operating costs. Therefore, suppressing shaft currents is crucial for reducing wind turbine costs.

[0038] The present invention addresses the path through which the generator shaft current flows. Insulated bearings and insulating end caps are used to alter the impedance of the shaft current path, blocking the shaft current path. By grounding the generator shaft at the drive and / or non-drive ends, a bypass is created for the shaft current, reducing the shaft current flowing through the bearings. Current transformers are installed on the shaft grounding cables at the drive and / or non-drive ends of the generator to monitor the shaft current. The measured values are sent to a data processing module, which analyzes and processes the data to form an assessment of the working status of the insulated bearings. This guides timely replacement of the insulating material of the insulated bearings, reduces operation and maintenance costs, and ensures equipment safety.

[0039] In a first aspect, a system for reducing generator shaft current damage is provided, wherein the generator main shaft is connected to the generator housing through at least one insulating bearing, and at least one grounding ring is provided on the generator main shaft; the grounding ring is grounded by sequentially connected carbon brushes and grounding wires; a current transformer is provided on the grounding wire, and the current transformer is used to obtain grounding current data and transmit it to a data processing end, such as Figure 2 As shown; the data processing end is used to predict the working status data of the insulating bearing based on the grounding current data, the environmental data of the fan, the fan operation data and the insulating bearing data, and evaluate whether to issue a shaft current risk alarm based on the insulating bearing working status data.

[0040] Among them, ground current data includes current values; fan environment data includes temperature, humidity, and salt spray concentration; fan operation data includes vibration signal data; and insulated bearing data includes insulation layer thickness, dielectric strength, and material aging coefficient. It can be seen that insulation layer thickness, dielectric strength, and material aging coefficient are fixed insulation material performance data. Once the insulation material for the bearing is selected, this performance data is essentially determined. Therefore, after the insulation material selection is completed, the insulation material performance data can be directly entered into the data processing terminal.

[0041] The main shaft of a wind turbine consists of two ends: one end connects to the blades and is therefore the driven end; the other end is the non-driven end. Since both ends of the main shaft extend through the generator casing, bearings secure the two ends at the point where the main shaft penetrates the generator casing, preventing them from interfering with the main shaft's rotation. When differential-mode shaft currents are generated, they can easily flow from one end of the main shaft through the bearing, the casing, the other end, and the bearing, generating a circulating current that can erode the bearing components. Due to the high cost of main shaft bearings, failure to limit shaft currents significantly shortens bearing life, leading to increased production costs.

[0042] In one embodiment of the present invention, the bearing at one end of the spindle (either the driven or non-driven end) can be replaced with an insulated bearing, and a grounding ring can be installed at the driven or non-driven end of the spindle. The grounding ring is connected to the ground via a carbon brush and a grounding wire, and a current transformer is installed on the grounding wire to collect real-time ground current data. In this embodiment, the insulated bearing at one end of the spindle prevents the formation of differential-mode shaft currents along the path from spindle to bearing to housing to the other end bearing and then to spindle.

[0043] Because there is only one ground terminal, high ground currents can cause insulation breakdown in the insulated bearing. Therefore, building on the above-mentioned implementation, a grounding ring can be installed at both the drive and non-drive ends of the spindle. The grounding ring is then connected to the ground via a carbon brush and a grounding wire, and a current transformer is installed on the grounding wire to collect real-time ground current data. Furthermore, by shunting the current between the two grounding paths, the grounding current at a single end is reduced, thereby minimizing the possibility of insulation breakdown.

[0044] In addition, in the present invention, a better embodiment is as follows Figure 1 As shown, the bearings on both the driven and non-driven ends of the main shaft are replaced with insulated bearings. Grounding rings are installed at both ends of the main shaft. These rings are connected to the ground via carbon brushes and grounding wires, and a current transformer is installed on the grounding wire to collect real-time ground current data. This not only shuns the ground current but also ensures that both ends of the main shaft are securely connected to the generator housing via the insulated bearings, further reducing the risk of insulation breakdown and extending the generator's operating life.

[0045] In the above embodiments, the current information introduced into the earth by the grounding wire can be obtained in real time through the current transformer arranged on the grounding wire; the acquisition of this current information can be used to monitor the working status information of the insulating material of the insulated bearing; in practice, when the working condition of the insulating material of the insulated bearing is poor, its impedance will decrease, generating a weaker shaft current, and the grounding current value will decrease at this time; that is to say, by real-time monitoring of the grounding current value, the working status information of the insulating material of the insulated bearing can be inferred to a certain extent; however, due to the relatively single information source, there is a greater possibility of misjudgment, so it is necessary to introduce more information to realize the evaluation of the working status of the insulating material of the insulated bearing, such as obtaining the performance data of the insulating material of the insulated bearing and the environmental data and vibration data of the wind turbine. Based on the above data, the status of the insulating material of the insulated bearing, such as aging, can be evaluated by multivariate features.

[0046] In order to predict the working status of the insulating material of the insulating bearing, a sensor module is also provided in the system, which includes at least a temperature and humidity sensor, a salt spray concentration sensor, and a vibration sensor; the temperature and humidity sensor, the salt spray concentration sensor, and the vibration sensor are all connected to the data processing end, and the required data information is summarized through the data processing end to realize the prediction of the working status of the insulating material of the insulating bearing based on artificial intelligence technology.

[0047] Specifically, a data processing terminal can be equipped with a built-in insulated bearing working state prediction model. By importing the real-time current value, temperature value, humidity value, salt spray concentration value, and vibration signal data into the insulated bearing working state prediction model, the current insulated bearing working state can be obtained. It should be noted that the data processing terminal can be a computer device deployed at the wind turbine. However, due to the limited computing power of the computer equipment on the wind turbine, the insulated bearing working state prediction model needs to be as small as possible. When constructing the bearing working state prediction model based on a regression task using a neural network model, the dimensionality of the model's hidden layer can be appropriately reduced to ensure its computing power is matched. This method may reduce the prediction accuracy.

[0048] At the same time, in order to improve the prediction accuracy, the data processing end in the present invention can also be set in the centralized control center of the site, and the relevant data obtained by the equipment in the wind turbine can be transmitted to the centralized control center through the wireless network. The centralized control center then uses the deployed insulating bearing working status prediction model to predict the working status of the insulating material of the wind turbine insulating bearing, and saves the results of each prediction for review by the operation and maintenance personnel.

[0049] In order to make the operation and maintenance of the wind turbine more automated, the present invention can further determine by computer whether the obtained prediction evaluation result of the working status of the insulating material is within the preset working status range of the insulating bearing. If so, it is evaluated as being in good condition; otherwise, it is evaluated as being in poor condition, and the shaft current risk alarm is activated to remind the operation and maintenance personnel to further determine whether the insulating material of the insulating bearing needs to be replaced.

[0050] The present invention utilizes insulating bearings and insulating end caps to alter the impedance of the shaft current path, blocking the shaft current path. This path is bypassed by grounding the shaft at the generator's drive and / or non-drive ends, thereby reducing the shaft current flowing through the bearings. Current transformers are installed on the shaft grounding cables at the generator's drive and / or non-drive ends to monitor the shaft current. The measured values are sent to a data processing terminal for analysis and processing to assess the operating status of the insulated bearings. This allows for timely replacement of the bearing's insulation material, reducing operational costs and ensuring equipment safety.

[0051] The above scheme mainly records the system structure and working principle of the present invention. On the other hand, the present invention also provides a working method for reducing generator shaft current damage, which is used in the above system for reducing generator shaft current damage. Figure 3 As shown, the working method is specifically as follows:

[0052] Real-time acquisition of ground current data, wind turbine environment data, and wind turbine operation data and preprocessing; the preprocessing specifically includes: cleaning and interpolation of abnormal values for current values, temperature values, humidity values, salt spray concentration values, and vibration signal data. For example, if the current value suddenly reaches a maximum or minimum value, it may be a sensor failure, which needs to be eliminated or corrected using statistical methods or rule-based methods; when data is missing due to sensor failure or time series data is missing after data elimination, it is necessary to use interpolation methods to interpolate to complete the time series data, such as using linear interpolation or front and back filling, or using a model to predict missing values. The above-mentioned processing methods are all mature technologies. The core point of the present invention is to use artificial intelligence technology to realize the prediction and evaluation of the working status of the insulating material of the insulating bearing of the wind turbine based on the proposed data.

[0053] The current fluctuation variance data is obtained based on the processed current values. The average temperature, humidity, and salt spray concentration values within the data collection period are obtained based on the temperature, humidity, and salt spray concentration values, respectively. The frequency domain characteristics of the vibration signal are obtained based on the processed vibration signal data. Since the current, temperature, humidity, salt spray concentration, and vibration signal data are all time series data, further processing is required to reduce the data processing workload. The frequency domain characteristics of the vibration signal data can be obtained using techniques such as Fourier transform and power spectral density.

[0054] The pre-processed data is input into the insulated bearing working state prediction model to obtain the current insulated bearing working state; wherein the insulated bearing working state prediction model is constructed using a two-layer fully connected neural network;

[0055] The dimension of the first layer of the neural network is 64 dimensions, and the expression is as follows:

[0056] h1=ReLU(W1x+b1);

[0057] The dimension of the second layer of the neural network is 32 dimensions, and the expression is as follows:

[0058] h2=ReLU(W2h1+b2);

[0059] The dimension of the output layer is 1, and the expression is as follows:

[0060] y=σ(W3h2+b3);

[0061] Where x is the input matrix, W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias vectors, and ReLU is the activation function. The x matrix includes current fluctuation variance data, temperature average, humidity average, salt spray concentration average, vibration signal frequency domain characteristics, insulation layer thickness, dielectric strength, and material aging coefficient, i.e., x = (x1, x2, x3, x4, x5, x6, x7, x8) T W1, W2, W3, b1, b2, and b3 are all parameters obtained through optimization and adjustment of the model training using a large amount of historical data.

[0062] It should be noted that the dimensions of each level in the above neural network model can be adjusted based on the different model deployment locations.

[0063] Since the above model is designed based on regression tasks, the loss function of the above model combines regression tasks with noise reduction requirements. Huber Loss is used as the loss function to balance the mean absolute error and mean square error. The expression of the loss function is as follows:

[0064]

[0065] Where y is the actual working state of the insulation material, y′ is the working state of the insulation material predicted by the model, and δ is a hyperparameter, which is usually set to 1.0 to control the smoothing interval of the loss function.

[0066] After obtaining the working status of the insulating material of the insulated bearing through the above-mentioned prediction model, it is evaluated whether it is necessary to issue a shaft current risk alarm based on the obtained insulating bearing working status data and the preset insulating bearing working status range to remind the operation and maintenance personnel to confirm whether the insulating material of the insulating bearing needs to be replaced.

[0067] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A system for reducing generator shaft current damage, characterized in that: The generator main shaft is connected to the generator housing through at least one insulating bearing, and at least one grounding ring is also provided on the generator main shaft; The grounding ring is grounded by sequentially connected carbon brushes and grounding wires; A current transformer is provided on the grounding wire, and the current transformer is used to obtain grounding current data and transmit it to the data processing end; The data processing end is used to predict the insulating bearing working status data based on the ground current data, the wind turbine environment data, the wind turbine operation data and the insulating bearing data, and to evaluate whether to issue a shaft current risk alarm based on the insulating bearing working status data.

2. The system for reducing generator shaft current damage according to claim 1, characterized in that: There are two insulating bearings, and the driving end and the non-driving end of the generator are both connected to the generator housing through the insulating bearings.

3. The system for reducing generator shaft current damage according to claim 1, characterized in that: There are two grounding rings, and both the driving end and the non-driving end of the generator are provided with the grounding ring.

4. The system for reducing generator shaft current damage according to claim 1, wherein: It also includes a sensor module, which includes a temperature and humidity sensor, a salt spray concentration sensor, and a vibration sensor; the temperature and humidity sensor, the salt spray concentration sensor, and the vibration sensor are all connected to the data processing end.

5. The system for reducing generator shaft current damage according to claim 4, characterized in that: The grounding current data includes the current value; the environmental data of the fan includes the temperature value, humidity value and salt spray concentration value; the fan operation data includes vibration signal data; the insulating bearing data includes the insulation layer thickness, dielectric strength and material aging coefficient.

6. The system for reducing generator shaft current damage according to claim 5, characterized in that: The data processing end has a built-in insulated bearing working state prediction model, which obtains the current working state of the insulated bearing by importing the real-time current value, temperature value, humidity value, salt spray concentration value, and vibration signal data into the insulated bearing working state prediction model.

7. The system for reducing generator shaft current damage according to claim 6, characterized in that: The working status assessment is that when the working status data of the insulating bearing is within a preset insulating bearing working status range, it is assessed as being in good condition; otherwise, it is assessed as being in poor condition, and a shaft current risk alarm is activated.

8. A method for reducing generator shaft current damage, used in the system for reducing generator shaft current damage according to any one of claims 1 to 7, characterized in that: include: Real-time acquisition of ground current data, wind turbine environment data, and wind turbine operation data and pre-processing; The pre-processed data is input into the insulated bearing working state prediction model to obtain the current insulated bearing working state; Based on the current insulated bearing working status data and the preset insulated bearing working status range, it is evaluated whether to issue a shaft current risk alarm.

9. The working method for reducing generator shaft current damage according to claim 8, characterized in that: The pretreatment includes: Perform outlier cleaning and interpolation processing on current values, temperature values, humidity values, salt spray concentration values and vibration signal data; The current fluctuation variance data is obtained based on the processed current value; the average temperature, average humidity and average salt spray concentration in the data collection period are obtained based on the temperature, humidity and salt spray concentration values; the frequency domain characteristics of the vibration signal are obtained based on the processed vibration signal data.

10. The working method for reducing generator shaft current damage according to claim 8, characterized in that: The insulating bearing working state prediction model is constructed using a double-layer fully connected neural network; The dimension of the first layer of the neural network is 64 dimensions, and the expression is as follows: h1=ReLU(W1x+b1); The dimension of the second layer of the neural network is 32 dimensions, and the expression is as follows: h2=ReLU(W2h1+b2); The dimension of the output layer is 1, and the expression is as follows: y=σ(W3h2+b3); Among them, x is the input matrix, W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias vectors, and ReLU is the activation function.

Images