System and method for monitoring oil return temperature of clutch of single-shaft combined cycle generator set

Abstract

The invention relates to the field of clutch oil return temperature monitoring, and provides a single-shaft combined cycle generator set clutch oil return temperature monitoring system and method, and the system comprises a data collection module which is used for collecting the real-time temperature of an oil return pipeline of a clutch in a single-shaft combined cycle generator set in real time, and carrying out the normalization processing of the real-time temperature to obtain a normalized data set; the topological graph generation module is used for constructing a clutch topological structure chart containing temperature monitoring points in combination with the coordinate parameters of the oil return pipeline, weighting the temperature monitoring points in the clutch topological structure chart to obtain an updated clutch topological structure chart, and outputting a pipeline temperature evaluation value in combination with a global detection algorithm; and the temperature analysis module is used for evaluating the state of the clutch by combining the transient torque, the steady-state torque, the load parameter and the pipeline temperature evaluation value of the generator set. The problem that an existing oil return temperature monitoring module can only monitor the oil return temperature of the clutch and cannot accurately judge the working state of an oil return pipeline of the clutch based on the oil return temperature is solved.

Description

Technical Field

[0001] This invention relates to the field of clutch return oil temperature monitoring technology for single-shaft combined cycle generator sets, and particularly to a clutch return oil temperature monitoring system and method for single-shaft combined cycle generator sets. Background Technology

[0002] As a key component, the clutch of a single-shaft combined cycle generator set directly affects the safe operation and economy of the unit. The clutch is primarily used to connect and disconnect the gas turbine and steam turbine. When needed, it allows the gas turbine and steam turbine to operate synchronously and generate electricity together; when not needed, they can be easily separated for individual maintenance or repair.

[0003] Chinese invention patent CN106641045B discloses a safety monitoring system and method for a synchronous clutch in a combined cycle power plant. The system includes a monitoring device and a diagnostic device. The monitoring device includes: a return oil temperature monitoring module for monitoring the return oil temperature during clutch operation; an intermediate component sliding process monitoring module for monitoring the sliding process of the intermediate component during clutch disengagement and engagement; and a torque monitoring module for measuring the torque transmitted by the clutch's driving component. The monitoring device outputs the monitored data to the diagnostic device, which diagnoses the clutch's operating status based on the data output by the monitoring device. However, the above system can only monitor the return oil temperature during clutch operation through the return oil temperature monitoring module and cannot accurately determine the operating status of the clutch's return oil pipeline based on the return oil temperature. Summary of the Invention

[0004] The present invention aims to solve at least one of the problems existing in the prior art, and provides a system and method for monitoring the return oil temperature of a single-shaft combined cycle generator set clutch.

[0005] In one aspect, the present invention provides a clutch return oil temperature monitoring system for a single-shaft combined cycle generator set, the monitoring system comprising:

[0006] The data acquisition module is used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set, and to normalize the real-time temperature of the return oil line to obtain a normalized dataset.

[0007] The topology generation module is used to construct a clutch topology diagram containing temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, retrieve the load curve of the single-shaft combined cycle generator set, assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve, obtain an updated clutch topology diagram, and output pipeline temperature evaluation values ​​based on the updated clutch topology diagram and a global detection algorithm.

[0008] The temperature analysis module is used to pre-build a condition monitoring model. Based on the condition monitoring model, combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, the module analyzes and evaluates the condition of the clutch and outputs the clutch condition assessment results.

[0009] Optionally, the monitoring system further includes:

[0010] The status warning module is used to acquire the clutch status assessment result in real time, determine whether the status assessment result meets the preset status threshold, and if not, trigger the clutch status warning command and generate a clutch status warning report.

[0011] The visualization module is communicatively connected to the topology generation module, the temperature analysis module, and the status warning module, respectively, and is used to visualize the clutch topology diagram, the clutch status evaluation results, and the clutch status warning report.

[0012] Optionally, the data acquisition module includes:

[0013] At least one set of data acquisition devices are deployed in the return oil line of the clutch in the single-shaft combined cycle generator set, and are used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set based on a preset sampling frequency.

[0014] The preprocessing unit is used to calculate the real-time temperature standard deviation based on the real-time temperature of the return oil pipeline;

[0015] The fluctuation judgment unit is used to preset the temperature fluctuation threshold and determine whether the real-time temperature standard deviation meets the temperature fluctuation threshold. If the real-time temperature standard deviation is less than the temperature fluctuation threshold, the oil return line of the clutch is considered to be stable and normal. If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, a temperature fluctuation command is triggered.

[0016] The real-time temperature standard deviation is calculated using the following formula:

[0017]

[0018] in, T represents the real-time temperature standard deviation. n This represents the real-time temperature of the return oil pipeline obtained from the nth sampling. Let represent the average real-time temperature of the return oil pipeline obtained from n samplings, α represent the local outlier factor, and t(n) represent the abnormal fluctuation function corresponding to the real-time temperature of the return oil pipeline obtained from the nth sampling. T max T min These represent the maximum and minimum temperature values ​​in the real-time temperature of the return oil pipeline, respectively, where N is the total number of temperature data points.

[0019] The normalization unit is used to normalize the real-time temperature of the return oil pipeline based on the real-time temperature standard deviation to obtain a normalized dataset.

[0020] Optionally, the topology graph generation module includes:

[0021] The topology building unit is used to construct a clutch topology diagram containing temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil line.

[0022] The monitoring point weighting unit is used to retrieve the load curve of the single-shaft combined cycle generator set, and assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain the updated clutch topology diagram.

[0023] The temperature assessment unit is used to output pipeline temperature assessment values ​​based on the updated clutch topology diagram and a global detection algorithm.

[0024] Optionally, the topology generation module is used to construct a clutch topology diagram including temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, retrieve the load curve of the single-shaft combined cycle generator set, and assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain an updated clutch topology diagram, including:

[0025] The topology building unit is used for:

[0026] Obtain the coordinate parameters of the return oil line of the clutch, use a graphical modeling tool to perform graphical modeling of the clutch, build a clutch oil flow diagram, and construct an initial topology diagram;

[0027] Several topological connection points are created based on the coordinate parameters. Convolution operation is performed on each topological connection point based on its attributes and features to obtain the topological weight vector of each topological connection point.

[0028] The topological weight vectors of each topological connection point are arranged in order of their respective numbers to obtain a topological self-matrix. The topological self-matrix is ​​then multiplied by itself to obtain a multi-level adjacency matrix. The multi-level adjacency matrix is ​​obtained by multiplying the adjacency matrices of multiple levels with their corresponding weights and then accumulating the results. The adjacency matrix of each level is obtained by multiplying the topological self-matrix of the same number as the level it belongs to. The weights of the adjacency matrices of lower levels are lower than the weights of the adjacency matrices of higher levels.

[0029] The linear regression function of the multi-level adjacency matrix is ​​calculated using a linear regression algorithm. Based on the historical temperature of each topological connection point, the average temperature and maximum temperature standard deviation of each topological connection point are obtained. The weight coefficient of each topological connection point is calculated by combining the linear regression function with the average temperature, maximum temperature standard deviation of each topological connection point and the load curve of the single-shaft combined cycle generator set.

[0030] The weight coefficients of each of the topological connection points are loaded, and each of the topological connection points is set as a temperature monitoring point. Based on the weight coefficients of each of the topological connection points, the temperature monitoring points are weighted to obtain the updated clutch topology diagram.

[0031] Optionally, the topology generation module is used to output pipeline temperature evaluation values ​​based on the updated clutch topology diagram combined with a global detection algorithm, including:

[0032] The topology graph generation module is used for:

[0033] Identify the temperature standard deviation and weighting coefficient of the temperature monitoring point in the clutch topology diagram, and retrieve the generator load associated with the temperature monitoring point in the load curve;

[0034] The temperature standard deviation, weighting coefficient, and generator load associated with the temperature monitoring point are used as inputs to the global detection algorithm, and the initial evaluation value is calculated using the global detection algorithm.

[0035] The initial evaluation value is calculated using the following formula:

[0036]

[0037] in, ω1 and ω2 represent the return oil pipeline influence factor and heterogeneity influence factor, respectively, β represents the weighting coefficient, Qt represents the generator load, ΔT represents the temperature difference at the sampling interval of temperature monitoring point j, and J represents the number of temperature monitoring points.

[0038] The initial evaluation value is integraled three times based on the gradient descent method. The three-dimensional integral value is then corrected using the Bayesian criterion based on the posterior probability of the three-dimensional integral, and the pipeline temperature evaluation value is output.

[0039] The pipeline temperature assessment value is calculated using the following formula:

[0040]

[0041] in, Let P1 represent the pipeline temperature assessment value at the j-th temperature monitoring point, l represent the number of features in the global detection algorithm, λ represent the error correction coefficient, and P1 represent the pipeline temperature assessment value at the j-th temperature monitoring point. t This represents the cubic integral value corresponding to the initial evaluation value, and

[0042] Optionally, the temperature analysis module is used to pre-build a condition monitoring model, including:

[0043] The temperature analysis module is used for:

[0044] The historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​of the generator sets are collected. The historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​are then subjected to dimensionality reduction and normalization processing to obtain at least one set of modeling samples.

[0045] Load at least one set of modeling samples, perform hybrid enhancement on the modeling samples using a hybrid enhancement technique, and divide the hybrid-enhanced modeling samples into a training set and a test set;

[0046] Using the VGG16 model as the initial model of the state monitoring model, a long short-term memory neural network and a Gaussian process regression algorithm are introduced into the initial model to obtain an initial state monitoring module, and a fusion experiment is conducted on the initial state monitoring model.

[0047] The initial state monitoring model is trained using the training set, and the weight coefficients are learned. Iterative training is performed based on a preset number of iterations to obtain a converged state monitoring model.

[0048] Load the test set, use the test set to test and verify the state monitoring model, obtain the test results, determine whether the test results meet the preset accuracy threshold, and if the test results meet the accuracy threshold, output the converged state monitoring model.

[0049] Optionally, the temperature analysis module is used to analyze and evaluate the state of the clutch based on the condition monitoring model combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, and output the clutch state assessment results, including:

[0050] The temperature analysis module is used for:

[0051] The transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set are acquired in real time.

[0052] The transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set are adaptively filtered and convolutionally fused using the state monitoring model to obtain a convolutionally fused feature vector.

[0053] The convolutional fusion feature vector is probabilistically thickened to increase its distribution density, generating a state importance distribution, and a probabilistic optimization method is used to predict the state evaluation value at the next time point.

[0054] The state evaluation value is calculated using the following formula:

[0055]

[0056] Where zx represents the state evaluation value, q z The probability augmentation value of the convolution-fused feature vector is represented by t, the sampling time period is t, and the state importance distribution coefficient is ε. i Let ut be the i-th input representation of the convolutionally fused feature vector, and let ut represent the state importance distribution function. Y represents the convolutional fusion feature difference vector, and ν represents the frequency of state importance distribution.

[0057] Another aspect of the present invention provides a method for monitoring the return oil temperature of a clutch in a single-shaft combined cycle generator set, the monitoring method comprising:

[0058] The real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set is collected in real time, and the real-time temperature of the return oil line is normalized to obtain a normalized dataset.

[0059] Based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, a clutch topology diagram containing temperature monitoring points is constructed. The load curve of the single-shaft combined cycle generator set is retrieved. The temperature monitoring points in the clutch topology diagram are weighted based on the load curve to obtain an updated clutch topology diagram. Based on the updated clutch topology diagram and the global detection algorithm, the pipeline temperature evaluation value is output.

[0060] A pre-built condition monitoring model is used to analyze and evaluate the condition of the clutch based on the condition monitoring model combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment value of the single-shaft combined cycle generator set, and output the clutch condition assessment result.

[0061] The clutch status assessment result is acquired in real time, and it is determined whether the status assessment result meets the preset status threshold. If not, a clutch status warning command is triggered and a clutch status warning report is generated.

[0062] The system visualizes the clutch topology diagram, the clutch status assessment results, and the clutch status early warning report.

[0063] Optionally, the real-time acquisition of the return oil line temperature of the clutch in the single-shaft combined cycle generator set, and the normalization processing of the real-time temperature of the return oil line to obtain a normalized dataset, specifically includes:

[0064] Based on a preset sampling frequency, the real-time temperature of the return oil pipeline of the clutch in the single-shaft combined cycle generator set is collected in real time.

[0065] Calculate the real-time temperature standard deviation based on the real-time temperature of the return oil pipeline;

[0066] A preset temperature fluctuation threshold is set, and it is determined whether the real-time temperature standard deviation meets the temperature fluctuation threshold. If the real-time temperature standard deviation is less than the temperature fluctuation threshold, the oil return line of the clutch is considered to be stable and normal. If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, a temperature fluctuation command is triggered.

[0067] Based on the real-time temperature standard deviation, the real-time temperature of the return oil pipeline is normalized to obtain the normalized dataset.

[0068] Compared with the prior art, the present invention solves the problem that the return oil temperature monitoring module in the existing system can only monitor the return oil temperature in the clutch working state and cannot accurately determine the working state of the clutch return oil pipeline based on the return oil temperature, and has the following beneficial effects.

[0069] 1. By constructing a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch return oil line, the system can more easily and accurately predict the pipeline temperature assessment value. Simultaneously, by combining the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of a single-shaft combined cycle generator set with a condition monitoring model, the system analyzes and evaluates the generator set clutch status. The pipeline temperature assessment value can more accurately determine the working state of the clutch return oil line, and the system can analyze the changing trend of the clutch's working state based on the pipeline temperature assessment value, predicting the future working state of the clutch. This overcomes the problem in existing systems where the return oil temperature monitoring module can only monitor the return oil temperature during clutch operation but cannot accurately determine the working state of the clutch return oil line based on the return oil temperature, nor can it assess the clutch's working state in conjunction with generator set operating parameters.

[0070] 2. The data acquisition module includes a data acquisition unit, a preprocessing unit, a fluctuation judgment unit, and a normalization unit. Through the coordinated operation of these units, anomaly prediction processing can be performed on the real-time temperature of the collected return oil line, thus achieving distributed processing of data acquisition and preprocessing. Simultaneously, by utilizing the fluctuation judgment unit to pre-evaluate the state of the clutch's return oil line, the system load and data processing volume are reduced, improving system smoothness.

[0071] 3. The clutch topology diagram provides a clear visual representation of the entire clutch return oil line system layout, ensuring temperature monitoring covers all critical components without omission. Furthermore, combined with a global detection algorithm, it can analyze and predict the temperature trend of the clutch return oil line, issuing a visual warning before temperature anomalies occur, effectively preventing damage to the clutch return oil line system.

[0072] 4. By using a global detection algorithm to fuse the temperature standard deviation, weighting coefficient, and generator load associated with the temperature monitoring points, and by using the gradient descent method to optimize the error of the initial evaluation value, the system can accurately predict and correct the pipeline temperature evaluation value, enabling the system to reflect the changes in the return oil pipeline temperature in real time.

[0073] 5. A condition monitoring model was constructed and trained. The condition monitoring model used the VGG16 model as the initial model and introduced a long short-term memory neural network and a Gaussian process regression algorithm. This model can realize the fusion analysis of multi-source data and effectively process image data or multi-dimensional sensor data. The Gaussian process regression algorithm can provide probabilistic predictions based on these features, thereby improving the robustness and accuracy of the model and achieving accurate assessment of the clutch status of a single-shaft combined cycle generator set. Attached Figure Description

[0074] One or more embodiments are illustrated by way of example with the corresponding pictures in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0075] Figure 1 A schematic diagram of a single-shaft combined cycle generator set clutch return oil temperature monitoring system provided in one embodiment of the present invention;

[0076] Figure 2 A flowchart illustrating the steps performed by the topology building unit according to another embodiment of the present invention;

[0077] Figure 3 A flowchart illustrating the steps performed by the topology graph generation module according to another embodiment of the present invention;

[0078] Figure 4 A flowchart illustrating a portion of the steps performed by the temperature analysis module according to another embodiment of the present invention;

[0079] Figure 5 A flowchart of another part of the steps performed by the temperature analysis module provided in another embodiment of the present invention;

[0080] Figure 6 A flowchart of a method for monitoring the return oil temperature of a clutch in a single-shaft combined cycle generator set, provided as another embodiment of the present invention;

[0081] Figure 7 A flowchart illustrating the specific steps included in step S10, which is provided for another embodiment of the present invention. Detailed Implementation

[0082] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the various embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and with various variations and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0083] One embodiment of the present invention relates to a clutch return oil temperature monitoring system for a single-shaft combined cycle generator set, the structure of which is as follows: Figure 1As shown, it includes a data acquisition module 100, a topology generation module 200, and a temperature analysis module 300.

[0084] The data acquisition module 100 is used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set, and to normalize the real-time temperature of the return oil line to obtain a normalized dataset.

[0085] The topology generation module 200 is used to construct a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch return oil pipeline. It retrieves the load curve of the single-shaft combined cycle generator set, assigns weights to the temperature monitoring points in the clutch topology diagram based on the load curve, and obtains an updated clutch topology diagram. Based on the updated clutch topology diagram and the global detection algorithm, it outputs pipeline temperature evaluation values.

[0086] The temperature analysis module 300 is used to pre-build a condition monitoring model. Based on the condition monitoring model and combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, the clutch condition is analyzed and evaluated, and the clutch condition assessment results are output.

[0087] Compared to existing technologies, this invention constructs a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch's return oil pipeline. This topology diagram allows for more convenient and accurate prediction of pipeline temperature assessment values. Furthermore, by combining a state monitoring model with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of a single-shaft combined cycle generator set, the clutch state is analyzed and evaluated. The pipeline temperature assessment values ​​allow for a more accurate determination of the clutch's return oil pipeline operating state and enable analysis of the clutch's operating state trends to predict future operating states. This overcomes the limitations of existing systems where the return oil temperature monitoring module can only monitor the return oil temperature during clutch operation but cannot accurately determine the clutch's return oil pipeline operating state based on the temperature, nor can it assess the clutch's operating state in conjunction with generator set operating parameters.

[0088] For example, such as Figure 1 As shown, the single-shaft combined cycle generator set clutch return oil temperature monitoring system also includes a status early warning module 400 and a visualization module 500.

[0089] The status warning module 400 is used to acquire the clutch status assessment results in real time, determine whether the status assessment results meet the preset status threshold, and if not, trigger the clutch status warning command and generate a clutch status warning report.

[0090] The visualization module 500 is communicatively connected to the topology generation module 200, the temperature analysis module 300, and the status warning module 400, respectively, and is used to visualize the clutch topology diagram, clutch status evaluation results, and clutch status warning reports.

[0091] Specifically, based on the clutch status warning command, the status warning module 400 can generate a clutch status warning report that includes the clutch status assessment results, the corresponding warning type, warning level, and warning handling measures, so as to provide timely and effective information to the relevant personnel and enable them to take appropriate handling measures in a timely manner.

[0092] The communication between the data acquisition module 100, the topology generation module 200, the temperature analysis module 300, the status early warning module 400, and the visualization module 500 can be achieved via Bluetooth, local area network, or 5G.

[0093] For example, such as Figure 1 As shown, the data acquisition module 100 includes at least one set of data acquisition units 110, a preprocessing unit 120, a fluctuation judgment unit 130, and a normalization unit 140.

[0094] The data acquisition unit 110 is deployed in the return oil line of the clutch in the single-shaft combined cycle generator set, and is used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set based on a preset sampling frequency.

[0095] The preprocessing unit 120 is used to calculate the real-time temperature standard deviation based on the real-time temperature of the return oil pipeline.

[0096] The fluctuation judgment unit 130 is used to preset the temperature fluctuation threshold and judge whether the real-time temperature standard deviation meets the temperature fluctuation threshold. If the real-time temperature standard deviation is less than the temperature fluctuation threshold, the return oil line of the clutch is considered to be stable and normal. If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, the temperature fluctuation command is triggered.

[0097] The real-time temperature standard deviation is calculated using the following formula:

[0098]

[0099] in, T represents the real-time temperature standard deviation. n This represents the real-time temperature of the return oil pipeline obtained from the nth sampling. Let represent the average real-time temperature of the return oil pipeline obtained from n samplings, α represent the local outlier factor, and t(n) represent the abnormal fluctuation function corresponding to the real-time temperature of the return oil pipeline obtained from the nth sampling. T max T minThese represent the maximum and minimum temperature values ​​in the real-time temperature data of the return oil pipeline, respectively. N is the total number of temperature data points, and also the maximum value of n.

[0100] For example, the temperature fluctuation threshold can be set to any value between 0.05 and 0.08. The local outlier factor can be set to any value between 0.001 and 0.003. The anomaly fluctuation function can be set to a one-class support vector machine (SVM) function.

[0101] The normalization unit 140 is used to normalize the real-time temperature of the return oil pipeline based on the real-time temperature standard deviation to obtain a normalized dataset.

[0102] Specifically, the preprocessing unit 120 can communicate with multiple data acquisition units 110 in a one-to-many manner. The number of channels of the data acquisition unit 110 can be single-channel, dual-channel, four-channel, or eight-channel. The interface type of the data acquisition unit 110 can be USB, RS232, RS485, or Ethernet.

[0103] By configuring a data acquisition module with a data acquisition unit, a preprocessing unit, a fluctuation judgment unit, and a normalization unit, the real-time temperature of the acquired return oil line can be pre-judged and processed for anomalies through the coordinated operation of these units. This achieves distributed processing of data acquisition and preprocessing. Simultaneously, by utilizing the fluctuation judgment unit to pre-evaluate the state of the clutch's return oil line, the system load and data processing volume are reduced, improving system smoothness.

[0104] For example, such as Figure 1 As shown, the topology generation module 200 includes a topology building unit 210, a monitoring point weighting unit 220, and a temperature evaluation unit 230.

[0105] The topology building unit 210 is used to construct a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch return oil line.

[0106] The monitoring point weighting unit 220 is used to retrieve the load curve of the single-shaft combined cycle generator set, and to assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain the updated clutch topology diagram.

[0107] Temperature assessment unit 230 is used to output pipeline temperature assessment values ​​based on the updated clutch topology diagram and a global detection algorithm.

[0108] This implementation utilizes a clutch topology diagram to visually display the layout of the entire clutch return oil pipeline system, ensuring that temperature monitoring covers all critical components without omission. Simultaneously, combined with a global detection algorithm, it can analyze and predict the temperature trend of the clutch return oil pipeline, thereby issuing a visual warning before temperature anomalies occur, effectively preventing damage to the clutch return oil pipeline system.

[0109] For example, the topology generation module is used to construct a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch return oil line. It then retrieves the load curve of a single-shaft combined cycle generator set and assigns weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain an updated clutch topology diagram. This includes: a topology construction unit performing steps S101 to S105. The following section combines... Figure 2 Steps S101 to S105 will be explained in detail.

[0110] Step S101: Obtain the coordinate parameters of the clutch return oil line, use a graphical modeling tool to perform graphical modeling of the clutch, build the clutch oil flow diagram, and construct the initial topology diagram.

[0111] For example, in step S101, when using graphical modeling tools to graphically model the clutch, ANSYS Fluent, Simulink, Autodesk Inventor, etc., can be used for visual modeling.

[0112] Step S102: Create several topological connection points based on coordinate parameters, and perform convolution operation on each topological connection point based on its attributes and features to obtain the topological weight vector of each topological connection point.

[0113] Step S103: Arrange the topological weight vectors of each topological connection point in order of their index to obtain a topological self-matrix. Multiply the topological self-matrix by itself to obtain a multi-level adjacency matrix. The multi-level adjacency matrix is ​​obtained by multiplying the adjacency matrices of multiple levels with their corresponding weights and then accumulating the results. Each level's adjacency matrix is ​​obtained by multiplying the topological self-matrix of the same number as its level. The weights of lower-level adjacency matrices are lower than the weights of higher-level adjacency matrices.

[0114] For example, assuming the topological self-matrix is ​​A, then the multilevel adjacency matrix A multi It can be represented as A multi =w1A+w2A 2 +w3A 3 +…+w k A k . Among them, A, A 2 A 3 Ak These are the adjacency matrices for levels 1, 2, 3, ..., k, respectively, w1, w2, w3, ..., w k A and A respectively 2 A 3 A k The corresponding weights, and w1 < w2 < w3 < ... < w k .

[0115] Step S104: Calculate the linear regression function of the multi-level adjacency matrix using a linear regression algorithm, and obtain the average temperature and maximum temperature standard deviation of each topological connection point based on the historical temperature of each topological connection point. Calculate the weight coefficient of each topological connection point by combining the linear regression function with the average temperature, maximum temperature standard deviation of each topological connection point and the load curve of the single-shaft combined cycle generator set.

[0116] Specifically, in step S104, when calculating the linear regression function of the multi-level adjacency matrix using the linear regression algorithm, the average temperature, maximum temperature standard deviation, and load curve of the single-shaft combined cycle generator unit corresponding to the topology connection point can be used as independent variables, and the weight coefficient of the topology connection point can be used as the dependent variable. The parameters in the linear regression function, i.e., the regression coefficients, are then solved using the least squares method or other optimization algorithms. These regression coefficients reflect the degree of influence of each independent variable on the dependent variable. After obtaining the parameters in the linear regression function, these parameters, along with the average temperature, maximum temperature standard deviation, and load curve of the single-shaft combined cycle generator unit corresponding to the topology connection point as independent variables, are substituted into the linear regression function to calculate the weight coefficient of the topology connection point as the dependent variable.

[0117] Step S105: Load the weight coefficients of each topology connection point, set each topology connection point as a temperature monitoring point, assign weights to each temperature monitoring point based on the weight coefficients of each topology connection point, and obtain the updated clutch topology diagram.

[0118] For example, the topology generation module is used to output pipeline temperature evaluation values ​​based on the updated clutch topology diagram combined with a global detection algorithm, including: the topology generation module is used to perform the following steps S201 to S203. The following is in conjunction with... Figure 3 Steps S201 to S203 will be explained in detail.

[0119] Step S201: Identify the temperature standard deviation and weighting coefficient of the temperature monitoring point in the clutch topology diagram, and retrieve the generator load associated with the temperature monitoring point in the load curve.

[0120] Step S202: The temperature standard deviation, weighting coefficient, and generator load associated with the temperature monitoring point are used as inputs to the global detection algorithm, and the initial evaluation value is calculated using the global detection algorithm.

[0121] The initial evaluation value is calculated using the following formula:

[0122]

[0123] in, The initial evaluation value is represented by ω1 and ω2, which represent the return oil pipeline influence factor and heterogeneity influence factor, respectively. β represents the weighting coefficient, Qt represents the generator load, and ΔT represents the temperature difference at the sampling interval of temperature monitoring point j. J represents the number of temperature monitoring points, which is also the maximum value of j.

[0124] Step S203: The initial evaluation value is integraled three times based on the gradient descent method. The posterior probability of the three-dimensional integral is used to correct the three-dimensional integral value corresponding to the initial evaluation value through the Bayesian criterion, and the pipeline temperature evaluation value is output.

[0125] The pipeline temperature assessment value is calculated using the following formula:

[0126]

[0127] in, Let P1 represent the pipeline temperature assessment value at the j-th temperature monitoring point, l represent the number of features in the global detection algorithm, λ represent the error correction coefficient, and P1 represent the value of the pipeline temperature assessment value at the j-th temperature monitoring point. t This represents the cubic integral value corresponding to the initial evaluation value, and

[0128]

[0129] By using a global detection algorithm to fuse the temperature standard deviation, weighting coefficient, and generator load associated with the temperature monitoring points, and by using the gradient descent method to optimize the error of the initial evaluation value, the system can accurately predict and correct the pipeline temperature evaluation value, enabling it to reflect the changes in the return oil pipeline temperature in real time.

[0130] For example, the temperature analysis module for pre-building a condition monitoring model includes: the temperature analysis module performing the following steps S301 to S307. The following is in conjunction with... Figure 4 Steps S301 to S307 will be explained in detail.

[0131] Step S301: Obtain the historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​of the generator set. Perform dimensionality reduction and normalization processing on the historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​of the generator set to obtain at least one set of modeling samples.

[0132] For example, step S301, which performs dimensionality reduction on historical generator transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values, can employ principal component analysis or linear discriminant analysis. The normalization of these historical generator transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​can be achieved using Z-score normalization.

[0133] Step S302: Load at least one set of modeling samples, perform hybrid enhancement on the modeling samples using hybrid enhancement technology, and divide the hybrid enhanced modeling samples into training set and test set.

[0134] Specifically, in step S302, when performing hybrid enhancement on the modeling samples using hybrid enhancement techniques, one or more of the following methods can be employed: data augmentation, feature combination, noise injection, and sample resampling. The ratio of the training set to the test set can be set to 7:3. For example, data augmentation generates new samples by randomly transforming the original data. For instance, for time series data, data augmentation can be achieved through operations such as time shifting, scaling, and stretching. For image or tabular data, data augmentation can be achieved through operations such as rotation, flipping, cropping, and color transformation.

[0135] Step S303: Using the VGG16 model as the initial model for the state monitoring model, a long short-term memory neural network and a Gaussian process regression algorithm are introduced into the initial model to obtain the initial state monitoring module, and a fusion experiment is conducted on the initial state monitoring model.

[0136] Step S304: Use the training set to train the network and learn the weight coefficients of the initial state monitoring model, and perform iterative training based on the preset iterative training rounds to obtain a converged state monitoring model.

[0137] Specifically, the preset number of iterative training rounds can be any number between 200 and 500. A fusion experiment on the initial model incorporating the Long Short-Term Memory Neural Network and the Gaussian Process Regression algorithm can integrate these two algorithms, thereby fully utilizing their respective advantages and improving the model's predictive performance.

[0138] Step S305: Load the test set, use the test set to test and verify the state monitoring model, and obtain the test results.

[0139] Step S306: Determine whether the test result meets the preset accuracy threshold. If the test result meets the accuracy threshold, proceed to step S307 and output the converged state monitoring model; otherwise, return to step S304.

[0140] This implementation constructs and trains a condition monitoring model. The condition monitoring model uses the VGG16 model as the initial model and introduces a long short-term memory neural network and a Gaussian process regression algorithm. This enables the fusion analysis of multi-source data and effectively processes image data or multi-dimensional sensor data. The Gaussian process regression algorithm can provide probabilistic predictions based on these features, thereby improving the robustness and accuracy of the model and achieving accurate assessment of the clutch status of a single-shaft combined cycle generator set.

[0141] For example, the temperature analysis module is used to analyze and evaluate the clutch status based on the condition monitoring model combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, and output the clutch status assessment results, including: the temperature analysis module is used to perform the following steps S401 to S403. The following is in conjunction with... Figure 5 Steps S401 to S403 will be explained in detail.

[0142] Step S401: Real-time acquisition of transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set.

[0143] Step S402: The transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set are adaptively filtered and convolutionally fused using the condition monitoring model to obtain the convolutionally fused feature vector.

[0144] Step S403: Probabilistically thicken the convolutional fusion feature vector to increase its distribution density, generate a state importance distribution, and use a probabilistic optimization method to predict the state evaluation value at the next time point.

[0145] The status assessment value is calculated using the following formula:

[0146]

[0147] Where zx represents the state evaluation value, q z The probability augmentation value of the convolution-fused feature vector is represented by t, the sampling time period is t, and the state importance distribution coefficient is ε. i Let ut be the i-th input representation of the convolutionally fused feature vector, and let ut represent the state importance distribution function. Y represents the convolutional fusion feature difference vector, and ν represents the frequency of state importance distribution.

[0148] Another embodiment of the present invention relates to a method for monitoring the return oil temperature of a clutch in a single-shaft combined cycle generator set, the process of which is as follows: Figure 6 As shown, it includes steps S10 to S50.

[0149] Step S10: Real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set is collected, and the real-time temperature of the return oil line is normalized to obtain a normalized dataset.

[0150] Step S20: Based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, construct a clutch topology diagram containing temperature monitoring points. Retrieve the load curve of the single-shaft combined cycle generator set. Assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain an updated clutch topology diagram. Based on the updated clutch topology diagram and the global detection algorithm, output the pipeline temperature evaluation value.

[0151] Step S30: Pre-build a condition monitoring model. Based on the condition monitoring model and combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, analyze and evaluate the condition of the clutch and output the clutch condition assessment results.

[0152] Step S40: Obtain the clutch status assessment result in real time, determine whether the status assessment result meets the preset status threshold, and if not, trigger the clutch status warning command and generate a clutch status warning report.

[0153] Step S50: Visualize the clutch topology diagram, clutch status assessment results, and clutch status warning report.

[0154] For example, step S10 specifically includes steps S501 to S506. The following is in conjunction with... Figure 7 Steps S501 to S506 will be explained in detail.

[0155] Step S501: Based on the preset sampling frequency, the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set is collected in real time.

[0156] Step S502: Load the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set, and calculate the real-time temperature standard deviation based on the real-time temperature of the return oil line.

[0157] Step S503: Preset temperature fluctuation threshold and determine whether the real-time temperature standard deviation meets the temperature fluctuation threshold.

[0158] Step S504: If the real-time temperature standard deviation is less than the temperature fluctuation threshold, then it is judged to be stable and normal, that is, the return oil line of the clutch is considered to be stable and normal.

[0159] Step S505: If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, a temperature fluctuation command is triggered.

[0160] Step S506: Based on the real-time temperature standard deviation, normalize the real-time temperature of the return oil pipeline to obtain a normalized dataset.

[0161] The method for monitoring the return oil temperature of the clutch of a single-shaft combined cycle generator set provided in this invention can be implemented by the system for monitoring the return oil temperature of the clutch of a single-shaft combined cycle generator set provided in this invention, and will not be described in detail here.

[0162] Compared to existing technologies, this invention constructs a clutch topology diagram containing temperature monitoring points based on a normalized dataset and the coordinate parameters of the clutch's return oil pipeline. This topology diagram allows for more convenient and accurate prediction of pipeline temperature assessment values. Furthermore, by combining a state monitoring model with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of a single-shaft combined cycle generator set, the clutch state is analyzed and evaluated. The pipeline temperature assessment values ​​allow for a more accurate determination of the clutch's return oil pipeline operating state and enable analysis of the clutch's operating state trends to predict future operating states. This overcomes the limitations of existing systems where the return oil temperature monitoring module can only monitor the return oil temperature during clutch operation but cannot accurately determine the clutch's return oil pipeline operating state based on the temperature, nor can it assess the clutch's operating state in conjunction with generator set operating parameters.

[0163] Another embodiment of the present invention relates to a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the single-shaft combined cycle generator set clutch return oil temperature monitoring method described in the above embodiments.

[0164] That is, those skilled in the art will understand that all or part of the steps in the methods described in the above embodiments can be implemented by instructing related hardware through program instructions. These program instructions are stored in a computer-readable storage medium and include several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0165] Another embodiment of the present invention relates to an electronic device, including a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, implements the single-shaft combined cycle generator set clutch return oil temperature monitoring method described in the above embodiment.

[0166] Memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the single-shaft combined cycle generator set clutch return oil temperature monitoring method in the above embodiments of the present invention. Memory may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function. The data storage area may store data created by using the single-shaft combined cycle generator set clutch return oil temperature monitoring method, etc. Furthermore, memory may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0167] Those skilled in the art will understand that the above embodiments are specific implementations of the present invention, and in practical applications, various changes can be made in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A clutch return oil temperature monitoring system for a single-shaft combined cycle generator set, characterized in that, The monitoring system includes: The data acquisition module is used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set, and to normalize the real-time temperature of the return oil line to obtain a normalized dataset. The topology generation module is used to construct a clutch topology diagram containing temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, retrieve the load curve of the single-shaft combined cycle generator set, assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve, obtain an updated clutch topology diagram, and output pipeline temperature evaluation values ​​based on the updated clutch topology diagram and a global detection algorithm. The temperature analysis module is used to pre-build a condition monitoring model. Based on the condition monitoring model, combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, the module analyzes and evaluates the condition of the clutch and outputs the clutch condition assessment results.

2. The monitoring system according to claim 1, characterized in that, The monitoring system also includes: The status warning module is used to acquire the clutch status assessment result in real time, determine whether the status assessment result meets the preset status threshold, and if not, trigger the clutch status warning command and generate a clutch status warning report. The visualization module is communicatively connected to the topology generation module, the temperature analysis module, and the status warning module, respectively, and is used to visualize the clutch topology diagram, the clutch status evaluation results, and the clutch status warning report.

3. The monitoring system according to claim 1, characterized in that, The data acquisition module includes: At least one set of data acquisition devices are deployed in the return oil line of the clutch in the single-shaft combined cycle generator set, and are used to collect the real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set based on a preset sampling frequency. The preprocessing unit is used to calculate the real-time temperature standard deviation based on the real-time temperature of the return oil pipeline; The fluctuation judgment unit is used to preset the temperature fluctuation threshold and determine whether the real-time temperature standard deviation meets the temperature fluctuation threshold. If the real-time temperature standard deviation is less than the temperature fluctuation threshold, the oil return line of the clutch is considered to be stable and normal. If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, a temperature fluctuation command is triggered. The real-time temperature standard deviation is calculated using the following formula: in, T represents the real-time temperature standard deviation. n This represents the real-time temperature of the return oil pipeline obtained from the nth sampling. Let represent the average real-time temperature of the return oil pipeline obtained from n samplings, α represent the local outlier factor, and t(n) represent the abnormal fluctuation function corresponding to the real-time temperature of the return oil pipeline obtained from the nth sampling. T max T min These represent the maximum and minimum temperature values ​​in the real-time temperature of the return oil pipeline, respectively, where N is the total number of temperature data points. The normalization unit is used to normalize the real-time temperature of the return oil pipeline based on the real-time temperature standard deviation to obtain a normalized dataset.

4. The monitoring system according to claim 3, characterized in that, The topology graph generation module includes: The topology building unit is used to construct a clutch topology diagram containing temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil line. The monitoring point weighting unit is used to retrieve the load curve of the single-shaft combined cycle generator set, and assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain the updated clutch topology diagram. The temperature assessment unit is used to output pipeline temperature assessment values ​​based on the updated clutch topology diagram and a global detection algorithm.

5. The monitoring system according to claim 4, characterized in that, The topology generation module is used to construct a clutch topology diagram including temperature monitoring points based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, retrieve the load curve of the single-shaft combined cycle generator set, and assign weights to the temperature monitoring points in the clutch topology diagram based on the load curve to obtain an updated clutch topology diagram, including: The topology building unit is used for: Obtain the coordinate parameters of the return oil line of the clutch, use a graphical modeling tool to perform graphical modeling of the clutch, build a clutch oil flow diagram, and construct an initial topology diagram; Several topological connection points are created based on the coordinate parameters. Convolution operation is performed on each topological connection point based on its attributes and features to obtain the topological weight vector of each topological connection point. The topological weight vectors of each topological connection point are arranged in order of their respective numbers to obtain a topological self-matrix. The topological self-matrix is ​​then multiplied by itself to obtain a multi-level adjacency matrix. The multi-level adjacency matrix is ​​obtained by multiplying the adjacency matrices of multiple levels with their corresponding weights and then accumulating the results. The adjacency matrix of each level is obtained by multiplying the topological self-matrix of the same number as the level it belongs to. The weights of the adjacency matrices of lower levels are lower than the weights of the adjacency matrices of higher levels. The linear regression function of the multi-level adjacency matrix is ​​calculated using a linear regression algorithm. Based on the historical temperature of each topological connection point, the average temperature and maximum temperature standard deviation of each topological connection point are obtained. The weight coefficient of each topological connection point is calculated by combining the linear regression function with the average temperature, maximum temperature standard deviation of each topological connection point and the load curve of the single-shaft combined cycle generator set. The weight coefficients of each of the topological connection points are loaded, and each of the topological connection points is set as a temperature monitoring point. Based on the weight coefficients of each of the topological connection points, the temperature monitoring points are weighted to obtain the updated clutch topology diagram.

6. The monitoring system according to claim 5, characterized in that, The topology generation module is used to output pipeline temperature evaluation values ​​based on the updated clutch topology diagram and a global detection algorithm, including: The topology graph generation module is used for: Identify the temperature standard deviation and weighting coefficient of the temperature monitoring point in the clutch topology diagram, and retrieve the generator load associated with the temperature monitoring point in the load curve; The temperature standard deviation, weighting coefficient, and generator load associated with the temperature monitoring point are used as inputs to the global detection algorithm, and the initial evaluation value is calculated using the global detection algorithm. The initial evaluation value is calculated using the following formula: in, ω1 and ω2 represent the return oil pipeline influence factor and heterogeneity influence factor, respectively, β represents the weighting coefficient, Qt represents the generator load, ΔT represents the temperature difference at the sampling interval of temperature monitoring point j, and J represents the number of temperature monitoring points. The initial evaluation value is integraled three times based on the gradient descent method. The three-dimensional integral value is then corrected using the Bayesian criterion based on the posterior probability of the three-dimensional integral, and the pipeline temperature evaluation value is output. The pipeline temperature assessment value is calculated using the following formula: in, Let λ represent the pipeline temperature assessment value at the j-th temperature monitoring point, l represent the feature number of the global detection algorithm, and λ represent the error correction coefficient. This represents the cubic integral value corresponding to the initial evaluation value, and 7. The monitoring system according to claim 6, characterized in that, The temperature analysis module is used to pre-build a condition monitoring model, including: The temperature analysis module is used for: The historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​of the generator sets are collected. The historical transient torque, historical steady-state torque, historical load parameters, and historical pipeline temperature assessment values ​​are then subjected to dimensionality reduction and normalization processing to obtain at least one set of modeling samples. Load at least one set of modeling samples, perform hybrid enhancement on the modeling samples using a hybrid enhancement technique, and divide the hybrid-enhanced modeling samples into a training set and a test set; Using the VGG16 model as the initial model of the state monitoring model, a long short-term memory neural network and a Gaussian process regression algorithm are introduced into the initial model to obtain an initial state monitoring module, and a fusion experiment is conducted on the initial state monitoring model. The initial state monitoring model is trained using the training set, and the weight coefficients are learned. Iterative training is performed based on a preset number of iterations to obtain a converged state monitoring model. Load the test set, use the test set to test and verify the state monitoring model, obtain the test results, determine whether the test results meet the preset accuracy threshold, and if the test results meet the accuracy threshold, output the converged state monitoring model.

8. The system according to claim 7, characterized in that, The temperature analysis module is used to analyze and evaluate the state of the clutch based on the condition monitoring model combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set, and outputs the clutch state assessment results, including: The temperature analysis module is used for: The transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set are acquired in real time. The transient torque, steady-state torque, load parameters, and pipeline temperature assessment values ​​of the single-shaft combined cycle generator set are adaptively filtered and convolutionally fused using the state monitoring model to obtain a convolutionally fused feature vector. The convolutional fusion feature vector is probabilistically thickened to increase its distribution density, generating a state importance distribution, and a probabilistic optimization method is used to predict the state evaluation value at the next time point. The state evaluation value is calculated using the following formula: Where zx represents the state evaluation value, q z The probability augmentation value of the convolution-fused feature vector is represented by t, the sampling time period is t, and the state importance distribution coefficient is ε. i Let ut be the i-th input representation of the convolutionally fused feature vector, and let ut represent the state importance distribution function. Y represents the convolutional fusion feature difference vector, and ν represents the frequency of state importance distribution.

9. A method for monitoring the return oil temperature of a clutch in a single-shaft combined cycle generator set, characterized in that, The monitoring method includes: The real-time temperature of the return oil line of the clutch in the single-shaft combined cycle generator set is collected in real time, and the real-time temperature of the return oil line is normalized to obtain a normalized dataset. Based on the normalized dataset and the coordinate parameters of the clutch return oil pipeline, a clutch topology diagram containing temperature monitoring points is constructed. The load curve of the single-shaft combined cycle generator set is retrieved. The temperature monitoring points in the clutch topology diagram are weighted based on the load curve to obtain an updated clutch topology diagram. Based on the updated clutch topology diagram and the global detection algorithm, the pipeline temperature evaluation value is output. A pre-built condition monitoring model is used to analyze and evaluate the condition of the clutch based on the condition monitoring model combined with the transient torque, steady-state torque, load parameters, and pipeline temperature assessment value of the single-shaft combined cycle generator set, and output the clutch condition assessment result. The clutch status assessment result is acquired in real time, and it is determined whether the status assessment result meets the preset status threshold. If not, a clutch status warning command is triggered and a clutch status warning report is generated. The system visualizes the clutch topology diagram, the clutch status assessment results, and the clutch status early warning report.

10. The monitoring method according to claim 9, characterized in that, The real-time temperature of the clutch return oil line in the single-shaft combined cycle generator set is collected in real time. The real-time temperature of the return oil line is normalized to obtain a normalized dataset, which specifically includes: Based on a preset sampling frequency, the real-time temperature of the return oil pipeline of the clutch in the single-shaft combined cycle generator set is collected in real time. Calculate the real-time temperature standard deviation based on the real-time temperature of the return oil pipeline; A preset temperature fluctuation threshold is set, and it is determined whether the real-time temperature standard deviation meets the temperature fluctuation threshold. If the real-time temperature standard deviation is less than the temperature fluctuation threshold, the oil return line of the clutch is considered to be stable and normal. If the real-time temperature standard deviation is greater than the temperature fluctuation threshold, a temperature fluctuation command is triggered. Based on the real-time temperature standard deviation, the real-time temperature of the return oil pipeline is normalized to obtain the normalized dataset.

Images