Automatic fault positioning method for flange fastening bolt of wind generating set

Abstract

The invention discloses an automatic fault positioning method for a fan flange fastening bolt. The method comprises the following specific implementation steps of: firstly, uniformly mounting ultrasonic sensors at a flange connecting part, and carrying out multi-point synchronous data acquisition; the method comprises the steps that axial stress data are preprocessed through a quartile method, the axial stress data of units of the same model in the initial operation state are determined as test data, the test data and real-time collected data are combined, an axial stress threshold value is set and optimized through a 3-sigma criterion and a dynamic adjustment coefficient, and preliminary abnormal state judgment is achieved. The method comprises the following steps: constructing a mechanical model based on the stress condition of a flange fastening bolt, obtaining the axial stress distribution characteristics of a bolt group under different working conditions by adopting a theoretical calculation method, optimizing an axial stress distribution model by utilizing test data, and introducing a multi-modal matrix for characterization so as to formulate a fault evaluation criterion. And finally, extracting stress distribution characteristics of the unmonitored bolts through a weight average method, and realizing fault positioning according to a fault evaluation criterion.

Description

Technical Field

[0001] The present invention belongs to the field of condition monitoring of wind turbine generators, and particularly relates to an automatic fault location method for flange fastening bolts of wind turbine generators. Background Art

[0002] With the continuous growth of the global demand for renewable energy, wind power generation, as a clean and efficient energy conversion method, has been widely applied. However, during the long-term operation of wind turbine generators, especially under harsh weather conditions, the reliability and safety of structural components face severe challenges. Among them, the flange fastening bolts of the wind turbine are important structural components connecting the blades, hub and tower barrel, and their condition monitoring and fault diagnosis are particularly important.

[0003] Traditional monitoring methods for flange fastening bolts of wind turbines mostly rely on manual inspection and regular maintenance, which are not only inefficient but also prone to undetected problems, making it difficult to timely discover potential faults. In addition, although current monitoring technologies are increasingly moving towards intelligence, due to the complex working conditions and diverse fault modes, accurate fault location remains a difficult problem.

[0004] Currently, the technologies applied to the field of condition monitoring of wind turbine generators mainly include methods such as acoustic monitoring, vibration analysis and temperature monitoring. However, these methods often cannot comprehensively and real-time reflect the stress conditions and potential faults of the bolts. Especially in the environment of large-scale wind farms, how to achieve the condition monitoring and fault location of the flange fastening bolts of the wind turbines urgently needs research and optimization. Summary of the Invention

[0005] The present invention belongs to the field of condition monitoring of wind turbine generators, and particularly relates to an automatic fault location method for flange fastening bolts of wind turbines. The specific content is as follows: First, ultrasonic sensors are installed at the flange fastening bolts in a uniformly distributed manner for multi-point synchronous acquisition. Data preprocessing is carried out based on the quartile algorithm. The axial stress data after preprocessing under the initial operating state is used as test data. By combining the test data and the real-time acquisition data, thresholds are set and adjusted based on the 3-σ criterion and the dynamic adjustment coefficient to achieve preliminary abnormal discrimination. Secondly, the stress condition of the flange fastening bolts is simplified into a mechanical model, and the axial stress distribution under different working conditions is obtained through theoretical calculation methods. Combining with the test data, the axial stress distribution model of the bolts is further optimized to obtain the fault location criterion. Finally, the eigenvalue of the unmonitored bolts is extracted by the weighted average method, and the bolt fault location is realized based on the fault location criterion. The overall framework of the present invention is as Figure 1 shown.

[0006] The purpose of the present invention is to provide an automatic fault location method for the fastening bolts of a fan flange, so as to solve the problems of low monitoring efficiency, high missed detection rate and inaccurate location in the prior art, thereby improving the operation safety and reliability of a wind power generation unit. The technology of the present invention mainly includes three modules: data acquisition, automatic anomaly discrimination and automatic fault location. The specific functions and implementation steps of each module are as follows:

[0007] The main function of the data acquisition module is to complete the tasks of data acquisition, storage and preprocessing. The specific implementation steps are as follows:

[0008] S101: Use an ultrasonic sensor to collect the ultrasonic echo signal of the axial stress of the bolt. After hardware and software filtering algorithms, remove the clutter and accurately extract the ultrasonic echo characteristic values of the axial stress of the bolt, so as to calculate the real-time axial stress data of the bolt;

[0009] S102: Screen and preprocess the axial stress data obtained in S101 according to the quartile method to remove the interference data therein;

[0010] The main function of the automatic anomaly discrimination module is to set and adaptively adjust the axial stress threshold of the bolt based on the test data set and the real-time collected data, and use the optimized threshold to achieve preliminary anomaly discrimination. The specific implementation steps are as follows:

[0011] S201: Based on the real-time wind condition data collected by the SCADA system and the preprocessing method described in S102, extract and merge the axial stress data of the bolts of the same model unit in the initial operation state and under the same wind condition. After preprocessing the obtained axial stress data, use it as the test data set;

[0012] S202: Use the collected test data set to set the basic threshold according to the 3-σ criterion;

[0013] S203: Collect the axial stress data of each monitored bolt to generate an axial stress time series. According to the real-time collected data and the actual working condition requirements, introduce a threshold adjustment coefficient to adaptively adjust the axial stress threshold. If the flange fastening bolt is affected by a large load fluctuation, the threshold adjustment coefficient increases to enhance the adaptability of the threshold to the fluctuation. For a stable operation scenario, the threshold adjustment coefficient decreases to avoid excessive false alarms and achieve adaptive threshold adjustment.

[0014] The main functions of the automatic fault location module include: establishing an axial stress distribution model of the flange fastening bolt as a fault evaluation criterion, equivalently calculating the axial stress data of the unmonitored bolts by combining the weighted average method and the geometric position relationship between the bolts, and determining the axial stress threshold of the unmonitored bolts. Finally, locate the bolts with faults. The specific process is as Figure 2 shown. The implementation steps are as follows:

[0015] S301: Establish an axial stress distribution model based on the mechanical analysis method to obtain the bolt axial stress distribution models under different working conditions (normal, single-region loosening or fracture, and multi-region loosening or fracture);

[0016] S302: Characterize the stress distribution characteristics of the flange fastening bolt group under different working conditions based on the multi-modal matrix principle to obtain the corresponding bolt axial stress distribution matrix, i.e., the flange fastening bolt fault assessment criterion;

[0017] S303: Based on the real-time data, use the weighted average method and the geometric position relationship between bolts to automatically extract the estimated values of the axial stress of the unmonitored bolts;

[0018] S304: Obtain the axial stress thresholds of each bolt according to the threshold setting and optimization algorithm described in S4 and S5;

[0019] S305: Combine the axial stress data of each bolt, the axial stress threshold, and the fault assessment criterion to perform automatic fault location and determine the fault type.

[0020] In summary, the present invention realizes the whole-process and real-time monitoring of the state of the flange fastening bolts of the fan by using ultrasonic sensors, overcomes the limitations of the traditional method, uses the initial operating state and wind condition data set of the same type of unit as the test data set, provides the axial stress threshold of the bolts more in line with the actual operating conditions of the wind turbine, and combines the 3-σ criterion, the threshold adjustment coefficient, the mechanical model, and the multi-modal matrix to provide a more scientific and accurate fault assessment criterion, significantly improving the monitoring accuracy. The weighted average method is introduced to extract the characteristic values, effectively expanding the monitoring range and improving the ability to locate the faults of unmonitored bolts. After implementing the present invention, it is possible to monitor the state of the flange fastening bolts of the fan in real time, accurately locate the possible fault points, reduce the safety hazards and economic losses caused by bolt faults, improve the overall operating efficiency and reliability of the wind turbine generator set. Transferring the acquisition of bolt fault characteristics to the software algorithm level is the basis for realizing the automatic positioning of fault bolts, which helps to promote the automation and intelligentization of the flange fastening bolt state monitoring and health assessment system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG Figure 1 is the overall framework schematic diagram of the present invention;

[0022] FIG Figure 2 is the flowchart of the automatic anomaly discrimination module of the present invention;

[0023] FIG Figure 3 is the flowchart of the automatic fault location module of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] The data acquisition module includes three sub-modules: data acquisition, data storage, and data preprocessing. The data acquisition sub-module uses ultrasonic sensors to collect ultrasonic echo signals of the axial stress of bolts. Through hardware and software filtering algorithms, clutter is removed, and the ultrasonic echo characteristic values of the bolt pre-tightening force are accurately extracted, thereby calculating the real-time axial stress data of the bolts. The data storage sub-module stores the axial stress data of each monitoring point of the flange fastening bolts in the database through SQL database language. The data preprocessing sub-module screens and preprocesses the obtained axial stress data according to the quartile method, removes the interference data therein, and improves the signal-to-noise ratio of the data.

[0025] The automatic anomaly discrimination module includes three sub-modules: test data set, threshold adjustment, and anomaly judgment. The specific process is as Figure 2 shown. The test data set is collected from the initial operating state of the flange fastening bolt group and obtained after preprocessing. The threshold adjustment sub-module needs to be jointly completed by the test data set and the real-time collected data. Based on the 3-σ criterion and the characteristic values extracted by the threshold adjustment coefficient, automatic threshold adjustment is achieved. The anomaly detection sub-module completes the preliminary anomaly discrimination based on the adjusted threshold and the real-time axial stress data.

[0026] The automatic fault location module mainly includes four sub-modules: establishing an axial stress distribution model based on mechanical analysis methods, fault evaluation criteria, feature extraction of unmonitored bolts, fault location, and output conclusion. The specific process is as Figure 3 shown. The feature extraction sub-module of unmonitored bolts automatically extracts the estimated value of the axial stress of unmonitored bolts based on real-time data and the weighted average method. The sub-module of establishing an axial stress distribution model and fault evaluation criteria based on mechanical analysis methods obtains the corresponding axial stress distribution characteristics under different working conditions through the axial stress distribution model, characterizes the axial stress distribution characteristics of the flange fastening bolt group under different working conditions based on the multi-modal matrix, and uses it as the fault evaluation criteria. This criterion is the key to realizing automatic fault location.

[0027] First, preprocess the collected data set. Arrange the axial stress data set in ascending order, divide it into four equal parts, and calculate the quartiles. The quartiles are the data at the three node positions after being divided into four equal parts. The corresponding formula is as follows:

[0028]

[0029] In the formula

[0030] represents the first quartile of the data set of the jth measuring point;

[0031] represents the second quartile of the data set of the jth measuring point;

[0032] represents the third quartile of the data set of the j-th measurement point;

[0033] n represents the total number of data contained in the data set;

[0034] and respectively represent the ((n + 1) / 2)-th, n / 2-th, and ((n + 2) / 2)-th numbers of the data set of the j-th measurement point;

[0035] m is a positive integer, and its size is determined by the value of n;

[0036] and respectively represent the m-th, (m + 1)-th, (m + 2)-th, (3m + 1)-th, and (3m + 3)-th numbers of the data set of the j-th measurement point;

[0037] IQR represents the interquartile range of the data set;

[0038] represents the lower boundary of the normal values in the data set of the j-th measurement point;

[0039] represents the upper boundary of the normal values in the data set of the j-th measurement point;

[0040] Furthermore, based on the real-time wind condition data collected by the SCADA system and the preprocessing method, the bolt axial stress data of the same type of unit in the initial operating state and under the same wind condition after preprocessing are combined as the test data set. Based on the test data set and the real-time collected data, the basic threshold is set and adjusted, and the preliminary anomaly discrimination is carried out. The specific implementation steps are as follows:

[0041] Step 1: Using the collected test data set, set the basic threshold according to the 3-σ criterion, and the corresponding formula is as follows:

[0042]

[0043] In the formula

[0044] represents the test axial stress data set of the j-th measurement point, and n is the total number of data

[0045] μ j represents the mean value of the test data set of the j-th measurement point;

[0046] σ j represents the variance of the test data set of the j-th measurement point;

[0047] represents the basic threshold of the test data set of the j-th measurement point.

[0048] Step 2: Collect the axial stress data of each monitoring bolt and generate an axial stress time series where N is the number of sensors. Dynamically adjust the axial stress threshold according to the real-time collected data and combined with the actual working condition requirements. The corresponding formula is as follows:

[0049]

[0050] In the formula

[0051] represents the real-time average axial stress amplitude of the j-th monitoring bolt;

[0052] represents the real-time axial stress data set of the j-th measuring point;

[0053] α j represents the threshold adjustment coefficient of the j-th measuring point;

[0054] represents the maximum value of the axial stress data of the j-th measuring point;

[0055] represents the minimum value of the axial stress data of the j-th measuring point;

[0056] τ j (t) represents the adjusted threshold of the axial stress data of the bolt at the j-th measuring point at time t.

[0057] Furthermore, the key to fault location lies in determining the fault assessment criterion. Based on the mechanical analysis method, establish an axial stress distribution model to obtain the corresponding axial stress distribution characteristics under different working conditions. The corresponding model formula is as follows:

[0058]

[0059] In the formula

[0060] F z represents the axial load for flange fastening;

[0061] F A(Z)i represents the axial force of the i-th bolt under the action of F z ;

[0062] n s represents the number of single flange fastening bolts;

[0063] (x i ,y i ) represents the centroid coordinates of the flange fastening bolt group;

[0064] I xx represents the moment of inertia of the flange fastening bolt group relative to the flapping direction;

[0065] I yy represents the moment of inertia of the flange fastening bolt group relative to the flapping direction;

[0066] M x represents the bending moment of the flange fastening bolt group in the flapping direction;

[0067] M y represents the bending moment of the flange fastening bolt group in the lead-lag direction;

[0068] F A(M)i represents the axial force of the i-th bolt under the action of M x and M y ;

[0069] P i represents the pre-tightening force of the i-th bolt;

[0070] C ib represents the stiffness of the i-th bolt;

[0071] C im represents the stiffness of the connecting part of the i-th bolt;

[0072] F Ai represents the total axial force of the i-th bolt;

[0073] represents the remaining number of single flange fastening bolts under fracture conditions;

[0074] (x * , y * ) represents the centroid coordinates of the flange fastening bolt group under fracture conditions;

[0075] represents the moment of inertia of the flange fastening bolt group relative to the flapping direction under fracture conditions;

[0076] represents the moment of inertia of the flange fastening bolt group relative to the lead-lag direction;

[0077] M zx represents the additional flapping direction bending moment M z generated when the axial load F zx is applied due to eccentricity;

[0078] M zy represents the additional lead-lag direction bending moment M z generated when the axial load F zx is applied due to eccentricity;

[0079] represents the axial force of the i-th bolt under the action of F z under fracture conditions;

[0080] Denote the axial force of the \(i\)-th bolt under the fracture condition at \(M\) x and \(M\) y acting on it;

[0081] Denote the total axial force of the \(i\)-th bolt under the fracture condition;

[0082] Furthermore, based on the multi-modal matrix theory, the axial stress distribution characteristics of the flange fastening bolt group under different working conditions are characterized. Specifically, the rows of the multi-modal matrix represent different working conditions, and the columns represent the axial stress values of each bolt. The specific formula is as follows:

[0083]

[0084] In the formula

[0085] \(S\) represents the multi-modal matrix of the axial stress distribution characteristics of the flange fastening bolt group under different working conditions;

[0086] [\(M_1\cdots M\) k \(\cdots M\) L represents different working conditions during the operation of the flange fastening bolt group, and \(L\) represents the total number of working conditions;

[0087] represents the axial stress vector when the flange fastening bolt group operates under the \(k\)-th type of working condition.

[0088] Furthermore, based on real-time data and test data sets, the axial stress distribution characteristics of unmonitored bolts are automatically extracted by using the weighted average method and the geometric position relationship between bolts. The specific extraction algorithm is as follows:

[0089]

[0090] In the formula

[0091] \(\omega\) ij represents the correlation between sensor \(i\) and unmonitored bolt \(j\), that is, the weight;

[0092] \(d\) ij represents the Euclidean distance between sensor \(i\) and unmonitored bolt \(j\);

[0093] \(N\) represents the number of sensors for a single flange;

[0094] \(x\) i \((t)\) represents the measured value of the axial stress of the monitored bolt \(i\) at time \(t\);

[0095] represents the estimated value of the axial stress of the unmonitored bolt \(j\) at time \(t\).

[0096] Finally, combining the joint fault assessment criterion, the axial stress distribution characteristics of the unmonitored bolts, and the model for adjusting and setting thresholds, automatic fault location is achieved, and the implementation steps are as follows:

[0097] Step 1: Establish an axial stress distribution model based on the mechanical analysis method to obtain the bolt axial stress distribution models under different working conditions;

[0098] Step 2: Characterize the stress distribution characteristics of the flange fastening bolt group under different working conditions based on the multi-modal matrix principle to obtain the multi-modal matrix of the bolt axial stress distribution, and use this matrix as the fault assessment criterion for the flange fastening bolts;

[0099] Step 3: Based on the preprocessed axial stress data, use the weighted average method and the geometric position relationship between bolts to automatically extract the estimated values of the axial stress of the unmonitored bolts;

[0100] Step 4: Set and adaptively adjust the axial stress thresholds of each bolt according to the existing threshold setting and optimization model;

[0101] Step 5: Combine the axial stress data of each bolt, the axial stress threshold, and the fault assessment criterion to perform automatic fault location and determine the fault type.

Claims

1. An automatic fault location method for flange fastening bolts of a wind turbine generator, characterized in that, The automatic fault location method includes three main modules: data acquisition, automatic abnormality identification and automatic fault location. The data acquisition module includes three sub-modules: data acquisition, data storage and data preprocessing; the automatic abnormality identification module includes three sub-modules: initial operation state test data set, threshold adjustment and abnormality judgment; the automatic fault location module mainly includes five sub-modules: establishing an axial stress distribution model based on mechanical analysis method, extracting features of unmonitored bolts, fault assessment criteria, fault location and output conclusion.

2. The automatic fault location method for flange fastening bolts of a wind turbine generator set according to claim 1, the data acquisition module, characterized in that The following steps are involved: Step 1: Use an ultrasonic sensor to collect the ultrasonic echo signal of the bolt axial stress; Step 2: Use hardware and software filtering algorithms to remove clutter and accurately extract the ultrasonic echo characteristic value of the bolt axial stress, thereby calculating the real-time axial stress data of the bolt; Step 3: The axial stress data obtained in step 2 are screened and preprocessed according to the quartile method to remove the interfering data. The corresponding formula is as follows: In the formula represents the first quantile of the data set at the j-th measurement point; represents the second quantile of the data set of the j-th measurement point; represents the third quartile of the data set of the j-th measurement point; n represents the total number of data contained in the data set; and respectively represent the ((n + 1) / 2)-th, the (n / 2)-th, and the ((n + 2) / 2)-th numbers of the j-th measuring point data set; m is a positive integer, and its size is determined by the value of n; and the m-th, (m + 1)-th, (m + 2)-th, (3m + 1)-th, and (3m + 3)-th numbers of the j-th measuring point data set, respectively; IQR represents the interquartile range of the data set; Represents the lower boundary of the normal value in the j-th measurement point dataset; Represents the upper boundary of the normal value in the dataset of the j-th measurement point.

3. The automatic fault location method for the flange fastening bolts of a wind turbine generator set according to claim 1, in the abnormal discrimination module, the test data set and the threshold adjustment sub-module, characterized in that, The test data set submodule is completed based on the real-time wind condition data collected by the SCADA system and the preprocessing method described in claim 2. First, the wind speed and wind direction information is obtained through the data collected by the SCADA system, and the information is input into the flange fastening bolt monitoring system. Then, the bolt axial stress data collected in the same wind speed and wind direction range under the initial operation state of each unit in the wind farm is extracted, and the bolt axial stress data is preprocessed. Finally, the bolt axial stress data of the same type of units under the same wind conditions are merged, which is the test data set; the threshold adjustment submodule is completed based on the test data set and the real-time collected data. The threshold adjustment coefficient is extracted to realize adaptive threshold adjustment. The specific implementation steps are as follows: Step 1: Use the collected test data set to set the basic threshold according to the 3-σ criterion. The corresponding formula is as follows: In the formula Denote the test axial stress data set of the j-th measuring point; μ j represents the mean of the test data set at the j-th measurement point; σ j represents the variance of the test data set at the j-th measurement point; Represents the basic threshold of the test data set for the j-th measurement point. Step 2: Collect the axial stress data of each monitoring bolt to generate an axial stress time series Where N is the number of sensors. According to the real-time collected data and the actual working condition requirements, a threshold adjustment coefficient is introduced to adaptively adjust the axial stress threshold. If the flange fastening bolts are affected by large load fluctuations, the threshold adjustment coefficient increases to enhance the adaptability of the threshold to the fluctuations. For stable operation scenarios, the threshold adjustment coefficient decreases to avoid excessive false alarms and achieve adaptive threshold adjustment. The corresponding formula is as follows: In the formula Denote the real-time average axial stress amplitude of the j-th monitoring bolt; Denote the real-time axial stress data set of the j-th measuring point; α j represents the threshold adjustment coefficient of the j-th measurement point; x max represents the maximum value of the axial stress data at the j-th measurement point; x min represents the minimum value of the axial stress data at the j-th measurement point; τ j (t) represents the adjusted threshold of the axial stress data of the bolt at the j-th measuring point at time t.

4. For an automatic fault location method of flange fastening bolts of a wind turbine generator set according to claim 1, the key lies in determining the fault evaluation criterion. An axial stress distribution model is established based on the mechanical analysis method in the automatic fault location module, and it is characterized in that, According to different working conditions, the corresponding axial stress distribution model is obtained, and then the axial stress distribution characteristics of different working conditions are obtained; in order to characterize the stress distribution characteristics of the flange fastening bolt group under different working conditions, a multi-modal matrix is ​​introduced, and the multi-modal matrix is ​​used as the fault assessment criterion. The corresponding model formula is as follows: Establish the axial stress distribution model based on mechanical analysis method: In the formula F z Indicates the axial load for flange fastening; F A(Z)i represents the axial stress of the i-th bolt under the action of F z ; n s Indicates the number of single flange fastening bolts; (x i , y i ) represents the centroid coordinates of the flange fastening bolt group; I xx represents the moment of inertia of the flange fastening bolt group with respect to the flapping direction; I yy represents the moment of inertia of the flange fastening bolt group with respect to the yawing direction; M x represents the bending moment of the flange fastening bolt group in the flapping direction; M y represents the bending moment of the flange fastening bolt group in the flapping direction; F A(M)i denotes the axial stress of the i-th bolt under the action of M x and M y ; P i represents the pre-tightening force of the i-th bolt; C ib represents the stiffness of the i-th bolt; C im represents the stiffness of the i-th bolt connection part; F Ai represents the total axial stress of the i-th bolt; Indicates the remaining number of single flange fastening bolts under the fracture condition; (x * , y * ) represents the centroid coordinates of the flange fastening bolt group under the fracture condition; Denote the moment of inertia of the flange fastening bolt group relative to the flapping direction under the fracture condition; Indicates the moment of inertia of the flange fastening bolt group relative to the shimmy direction; M zx represents the additional flapping moment M generated when an axial load F is applied due to eccentricity z ; zx ; M zy Indicates the additional flapping direction bending moment M z generated when the shaft load F is applied due to eccentricity zx ; Denote the axial force of the i-th bolt under the action of F z during the fracture condition; Denote the axial force of the i-th bolt under the action of M x and M y during the fracture condition; Denote the total axial force of the i-th bolt under the fracture condition; The axial stress distribution characteristics of the flange fastening bolt group under different working conditions are characterized based on the multimodal matrix, which is characterized in that the rows of the multimodal matrix represent different working conditions and the columns represent the axial stress values ​​of each bolt: In the formula S represents the multi-modal matrix of the axial stress distribution characteristics of the flange fastening bolt group under different working conditions; [M1...M k ...M L represents different operating conditions of the flange fastening bolt group during operation, and L represents the total number of operating conditions; Denote the axial stress vector when the flange fastening bolt group operates under the k-th type of working conditions.

5. The automatic fault location method for the flange fastening bolts of a wind turbine generator set according to claim 1, wherein in the automatic fault location module, there is no bolt feature extraction sub-module for monitoring, and it is characterized in that Based on real-time data and test data sets, the weighted average method and the geometric position relationship between bolts are used to automatically extract the axial stress distribution characteristics of unmonitored bolts. The corresponding extraction algorithm is as follows: In the formula ω ij Indicates the correlation between sensor i and the unmonitored bolt j, i.e., the weight; d ij denotes the Euclidean distance between sensor i and the unmonitored bolt j; N represents the number of sensors on a single blade; x i (t) represents the measured axial stress value of the monitored bolt i at time t; It represents the estimated value of the axial stress of bolt j not monitored at time t.

6. The automatic fault location method for the flange fastening bolts of a wind turbine generator set according to claim 1, characterized in that, The specific process of realizing the automatic fault location module function is as follows: Step 1: Establish an axial stress distribution model based on the mechanical analysis method according to Claim 4, and obtain the bolt axial stress distribution model under different working conditions; Step 2: Characterize the stress distribution characteristics of the flange fastening bolt group under different working conditions based on the multi-modal matrix according to Claim 4, obtain the bolt axial stress distribution multi-modal matrix, and use this matrix as the flange fastening bolt fault evaluation criterion; Step 3: Calculate the axial stress data of the unmonitored bolts according to the unmonitored bolt feature extraction algorithm described in Claim 5, and then obtain the axial stress data set of all bolts; Step 4: Set and adaptively adjust the axial stress thresholds of each bolt according to the threshold adjustment algorithm described in Claim 3; Step 5: Combine the axial stress data of each bolt, the axial stress threshold, and the fault evaluation criterion to perform automatic fault location and determine the fault type.

Images